Get the most recent info and news about WARDA on HackerNoon, where 10k+ technologists publish stories for 4M+ monthly readers.

We are problem solvers embarked on a journey to solve the digital chaos in fashion, retail, and luxury. We are the enabler of the digital transformation.

WARDA

warda

The attack also extended to the gathering of its vehicles and forces, located in Khalat Warda. Hezbollah utilized artillery shells and appropriate missiles during the operation, resulting in verified casualties. This incident underscores the ongoing ...

Hezbollah strikes Israel's 'Za'reit' battalion headquarters in Khalat Warda

Inspired by the remarkable durability of ancient Roman construction materials in seawater, a University of Texas at Arlington civil engineering researcher is attempting to duplicate Roman concrete by developing 3D-printed materials to restore damaged or dying coral reefs.

A bold plan to 3D print artificial coral reef

EXTRA a keda funda 1 di mart 1976, ku su meta prinsipal pa trese informashonnan alkaso, ophetivo, verifiká i sin sensashon negativo. Pa EXTRA edukashon di nos pueblo ta hopi haltu den nos prioridatnan.

Van Huffelen asombrá ku a warda te den pitu pa kaba di atendé refinansiamentu

FSRA has fined an auto body shop $75,000 and one of its then-employees $15,000 in connection with fraudulent business practices.

Regulator imposes hefty penalties in auto repair fraud case

Ontario’s insurance regulator has imposed administrative penalties and a compliance order against the auto repair shop McLaren Collision and an employee named Fady Rony Warda. A statement from the Financial Services Regulatory Authority of Ontario (FSRA) said the penalties are in response to fraudulent billing practices.

Auto repair shop faces heavy penalties over fraudulent billing practices

The Financial Services Regulatory Authority of Ontario (FSRA) has imposed a compliance order and an order imposing administrative penalties of $75,000 on McLaren Collision (McLaren) and $15,000 on Fady Warda (Warda).

FSRA imposes administrative penalties and compliance order against 1843538 Ontario Inc., carrying on business as McLaren Collision, and Fady Warda

Her uncle, his wife Soha, their five children Ahmed, 25, Marwan, 22, Asad, 10, Mohamed, 20, and Yazan, 10, plus Ahmed's wife Warda, 25, their two young boys, Abdel Naser, aged two, and two-month-old Omar, were all killed.

Gaza airstrike kills ten members of one family - including two-month-old baby

This paper is available on arxiv under CC 4.0 license. Authors: (1) Ali Ghanbari, Dept. of Computer Science, Iowa State University; (2) Deepak-George Thomas, Dept. of Computer Science, Iowa State University; (3) Muhammad Arbab Arshad, Dept. of Computer Science, Iowa State University; (4) Hridesh Rajan, Dept. of Computer Science, Iowa State University. Table of Links Abstract & Introduction Background Motivating Example Proposed Approach Supported DNN Bugs Evaluation Threats to Validity Related Work Conclusion, Acknowledgments & References IX. CONCLUSION This paper revisits mutation-based fault localization in the context of DNN and presents a novel DNN fault localization technique, named deepmufl. The technique is based on the idea of mutating a pre-trained DNN model and calculating suspiciousness values according to Metallaxis and MUSE approaches, Ochiai and SBI formulas, and two types of impacts of mutations on the results of test data points. deepmufl is compared to state-of-the-art static and dynamic fault localization systems [11], [8], [34], [12] on a benchmark of 109 model bugs. In this benchmark, while deepmufl is slower than the other tools, it proved to be almost two times more effective than them in terms of the total number of bugs detected and it detects 21 bugs that none of the studied tools were able to detect. We further studied the impact of mutation selection on fault localization time. We observed that we can halve the time taken to perform fault localization by deepmufl, while losing only 7.55% of the previously detected bugs. ACKNOWLEDGMENTS The authors thank Anonymous ASE 2023 Reviewers for their valuable feedback. We also thank Mohammad Wardat for his instructions on querying StackOverflow. This material is based upon work supported by the National Science Foundation (NSF) under the grant #2127309 to the Computing Research Association for the CIFellows Project. This work is also partially supported by the NSF grants #2223812, #2120448, and #1934884. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF. REFERENCES [1] IEEE Standard Classification for Software Anomalies, 2010. [2] A. McPeak, “What’s the true cost of a software bug?” https://smartbear. com/blog/software-bug-cost/, 2017, accessed 08/10/23. [3] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, no. 7553, pp. 436–444, 2015. [4] M. J. Islam, G. Nguyen, R. Pan, and H. Rajan, “A comprehensive study on deep learning bug characteristics,” in ESEC/FSE, 2019, pp. 510–520. [5] M. J. Islam, R. Pan, G. Nguyen, and H. Rajan, “Repairing deep neural networks: Fix patterns and challenges,” in ICSE, 2020, pp. 1135–1146. [6] Y. Zhang, Y. Chen, S.-C. Cheung, Y. Xiong, and L. Zhang, “An empirical study on tensorflow program bugs,” in ISSTA, 2018, pp. 129–140. [7] N. Humbatova, G. Jahangirova, G. Bavota, V. Riccio, A. Stocco, and P. Tonella, “Taxonomy of real faults in deep learning systems,” in ICSE, 2020, pp. 1110–1121. [8] M. Wardat, B. D. Cruz, W. Le, and H. Rajan, “Deepdiagnosis: automatically diagnosing faults and recommending actionable fixes in deep learning programs,” in ICSE. IEEE, 2022, pp. 561–572. [9] K. Pei, Y. Cao, J. Yang, and S. Jana, “Deepxplore: Automated whitebox testing of deep learning systems,” in SOSP, 2017, pp. 1–18. [10] J. Kim, R. Feldt, and S. Yoo, “Guiding deep learning system testing using surprise adequacy,” in ICSE, 2019, pp. 1039–1049. [11] M. Wardat, W. Le, and H. Rajan, “Deeplocalize: fault localization for deep neural networks,” in ICSE, 2021, pp. 251–262. [12] A. Nikanjam, H. B. Braiek, M. M. Morovati, and F. Khomh, “Automatic fault detection for deep learning programs using graph transformations,” TOSEM, vol. 31, no. 1, pp. 1–27, 2021. [13] M. Usman, D. Gopinath, Y. Sun, Y. Noller, and C. S. Pas˘ areanu, “Nn ˘ repair: Constraint-based repair of neural network classifiers,” in CAV, 2021, pp. 3–25. [14] X. Zhang, J. Zhai, S. Ma, and C. Shen, “Autotrainer: An automatic dnn training problem detection and repair system,” in ICSE, 2021, pp. 359–371. [15] W. E. Wong, R. Gao, Y. Li, R. Abreu, and F. Wotawa, “A survey on software fault localization,” TSE, vol. 42, no. 8, pp. 707–740, 2016. [16] M. Papadakis and Y. Le Traon, “Using mutants to locate” unknown” faults,” in ICST, 2012, pp. 691–700. [17] S. Moon, Y. Kim, M. Kim, and S. Yoo, “Ask the mutants: Mutating faulty programs for fault localization,” in ICST, 2014, pp. 153–162. [18] R. A. DeMillo, R. J. Lipton, and F. G. Sayward, “Hints on test data selection: Help for the practicing programmer,” IEEE Computer, vol. 11, pp. 34–41, 1978. [19] R. Abreu, P. Zoeteweij, and A. J. Van Gemund, “On the accuracy of spectrum-based fault localization,” in TAICPART-MUTATION, 2007, pp. 89–98. [20] V. Dallmeier, C. Lindig, and A. Zeller, “Lightweight Bug Localization with AMPLE,” in ISAADD, 2005, pp. 99–104. [21] J. A. Jones, M. J. Harrold, and J. Stasko, “Visualization of test information to assist fault localization,” in ICSE, 2002, pp. 467–477. [22] L. Naish, H. J. Lee, and K. Ramamohanarao, “A model for spectra-based software diagnosis,” TOSEM, vol. 20, no. 3, pp. 1–32, 2011. [23] S. Yoo, “Evolving human competitive spectra-based fault localisation techniques,” in SBSE, 2012, pp. 244–258. [24] X. Xie, F.-C. Kuo, T. Y. Chen, S. Yoo, and M. Harman, “Provably optimal and human-competitive results in sbse for spectrum based fault localisation,” in SBSE, 2013, pp. 224–238. [25] S. Ma, Y. Liu, W.-C. Lee, X. Zhang, and A. Grama, “Mode: automated neural network model debugging via state differential analysis and input selection,” in ESEC/FSE, 2018, pp. 175–186. [26] H. F. Eniser, S. Gerasimou, and A. Sen, “Deepfault: Fault localization for deep neural networks,” in FASE, 2019, pp. 171–191. [27] N. Humbatova, G. Jahangirova, and P. Tonella, “Deepcrime: mutation testing of deep learning systems based on real faults,” in ISSTA, 2021, pp. 67–78. [28] Q. Hu, L. Ma, X. Xie, B. Yu, Y. Liu, and J. Zhao, “Deepmutation++: A mutation testing framework for deep learning systems,” in ASE, 2019, pp. 1158–1161. [29] L. Ma, F. Zhang, J. Sun, M. Xue, B. Li, F. Juefei-Xu, C. Xie, L. Li, Y. Liu, J. Zhao et al., “Deepmutation: Mutation testing of deep learning systems,” in ISSRE, 2018, pp. 100–111. [30] M. Papadakis and Y. Le Traon, “Metallaxis-fl: mutation-based fault localization,” STVR, vol. 25, no. 5-7, pp. 605–628, 2015. [31] F. Chollet et al., “Keras,” https://keras.io, 2015. [32] M. Abadi, P. Barham, J. Chen, Z. Chen, A. Davis, J. Dean, M. Devin, S. Ghemawat, G. Irving, M. Isard, M. Kudlur, J. Levenberg, R. Monga, S. Moore, D. G. Murray, B. Steiner, P. Tucker, V. Vasudevan, P. Warden, M. Wicke, Y. Yu, and X. Zheng, “Tensorflow: A system for large-scale machine learning,” in OSDI, 2016, p. 265–283. [33] A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Kopf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala, “Pytorch: An imperative style, highperformance deep learning library,” in NIPS, 2019, pp. 8024–8035. [34] E. Schoop, F. Huang, and B. Hartmann, “Umlaut: Debugging deep learning programs using program structure and model behavior,” in CHI, 2021, pp. 1–16. [35] W. E. Wong and A. P. Mathur, “Reducing the cost of mutation testing: An empirical study,” JSS, pp. 185–196, 1995. [36] A. Ghanbari, D.-G. Thomas, M. A. Arshad, and H. Rajan, “Mutationbased fault localization of deep neural networks,” https://github.com/ ali-ghanbari/deepmufl-ase-2023, 2023. 37] J. H. Andrews, L. C. Briand, and Y. Labiche, “Is mutation an appropriate tool for testing experiments?” in ICSE, 2005, pp. 402–411. [38] M. Papadakis, M. Kintis, J. Zhang, Y. Jia, Y. Le Traon, and M. Harman, “Mutation testing advances: an analysis and survey,” in Advances in Computers, 2019, vol. 112, pp. 275–378. [39] V. Debroy and W. E. Wong, “Using mutation to automatically suggest fixes for faulty programs,” in ICST, 2010, pp. 65–74. [40] A. Ghanbari, S. Benton, and L. Zhang, “Practical program repair via bytecode mutation,” in ISSTA, 2019, pp. 19–30. [41] G. Fraser and A. Arcuri, “Achieving scalable mutation-based generation of whole test suites,” ESE, pp. 783–812, 2015. [42] F. C. M. Souza, M. Papadakis, Y. Le Traon, and M. E. Delamaro, “Strong mutation-based test data generation using hill climbing,” in IWSBST, 2016, pp. 45–54. [43] D. Shin, S. Yoo, M. Papadakis, and D.-H. Bae, “Empirical evaluation of mutation-based test case prioritization techniques,” STVR, p. e1695, 2019. [44] J. P. Galeotti, C. A. Furia, E. May, G. Fraser, and A. Zeller, “Inferring loop invariants by mutation, dynamic analysis, and static checking,” TSE, pp. 1019–1037, 2015. [45] A. Groce, I. Ahmed, C. Jensen, and P. E. McKenney, “How verified is my code? falsification-driven verification (t),” in ASE, 2015, pp. 737– 748. [46] B. Liblit, M. Naik, A. X. Zheng, A. Aiken, and M. I. Jordan, “Scalable statistical bug isolation,” ACM SIGPLAN Notices, vol. 40, no. 6, pp. 15–26, 2005. [47] R. Abreu, P. Zoeteweij, and A. J. Van Gemund, “An evaluation of similarity coefficients for software fault localization,” in PRDC, 2006, pp. 39–46. [48] Wikipedia contributors, “Hierarchical data format — Wikipedia, the free encyclopedia,” 2022, accessed 08/10/23. [49] H. Coles, T. Laurent, C. Henard, M. Papadakis, and A. Ventresque, “Pit: a practical mutation testing tool for java,” in ISSTA, 2016, pp. 449–452. [50] P. Ammann and J. Offutt, Introduction to software testing. Cambridge University Press, 2016. [51] R. Just, F. Schweiggert, and G. M. Kapfhammer, “Major: An efficient and extensible tool for mutation analysis in a java compiler,” in ASE, 2011, pp. 612–615. [52] Y.-S. Ma, J. Offutt, and Y. R. Kwon, “Mujava: an automated class mutation system,” STVR, pp. 97–133, 2005. [53] D. Schuler and A. Zeller, “Javalanche: Efficient mutation testing for java,” in ESEC/FSE, 2009, pp. 297–298. [54] “Junit,” http://junit.org/, 2019, accessed 08/10/23. [55] “Testng documentation,” https://testng.org/doc/documentation-main. html, 2017, accessed 08/10/23. [56] M. M. Morovati, A. Nikanjam, F. Khomh, Z. Ming et al., “Bugs in machine learning-based systems: A faultload benchmark,” arXiv, 2022. [57] X. Li and L. Zhang, “Transforming programs and tests in tandem for fault localization,” Proceedings of the ACM on Programming Languages, vol. 1, no. OOPSLA, pp. 1–30, 2017. [58] C. Parnin and A. Orso, “Are automated debugging techniques actually helping programmers?” in ISSTA, 2011, pp. 199–209. [59] P. S. Kochhar, X. Xia, D. Lo, and S. Li, “Practitioners’ expectations on automated fault localization,” in ISSTA, 2016, pp. 165–176. [60] D. Shin and D.-H. Bae, “A theoretical framework for understanding mutation-based testing methods,” in ICST, 2016, pp. 299–308. [61] A. V. Pizzoleto, F. C. Ferrari, J. Offutt, L. Fernandes, and M. Ribeiro, “A systematic literature review of techniques and metrics to reduce the cost of mutation testing,” Journal of Systems and Software, vol. 157, p. 110388, 2019. [62] J. Zhang, “Scalability studies on selective mutation testing,” in ICSE, vol. 2, 2015, pp. 851–854. [63] A. J. Offutt, J. Pan, K. Tewary, and T. Zhang, “An experimental evaluation of data flow and mutation testing,” Software: Practice and Experience, vol. 26, no. 2, pp. 165–176, 1996. [64] scikit-learn Contributors, “scikit-learn: Machine learning in python,” 2020, accessed 08/10/23. [Online]. Available: https://scikit-learn.org/stable/ [65] R. Heckel, “Graph transformation in a nutshell,” ENTCS, vol. 148, no. 1, pp. 187–198, 2006. [66] J. Cao, M. Li, X. Chen, M. Wen, Y. Tian, B. Wu, and S.-C. Cheung, “Deepfd: Automated fault diagnosis and localization for deep learning programs,” in ICSE, 2022, p. 573–585. [67] Y. Ishimoto, M. Kondo, N. Ubayashi, and Y. Kamei, “Pafl: Probabilistic automaton-based fault localization for recurrent neural networks,” IST, vol. 155, p. 107117, 2023. [68] Y. Sun, H. Chockler, X. Huang, and D. Kroening, “Explaining image classifiers using statistical fault localization,” in ECCV, 2020, pp. 391–406.

Mutation-based Fault Localization of Deep Neural Networks: Conclusion, Acknowledgments & References

DevOps for Scientific Computing is the alchemical fusion of containers, Git Repositories, and CI/CD Engines, birthing a revolution in scientific inquiry, where reproducibility and transparency meet the boundless frontiers of rigorous research. 🚀 Amid the dynamic realm of scientific inquiry, where curiosity fuels innovation and exploration knows no bounds, a groundbreaking transformation is underway. It's a transformation that converges the tenets of DevOps with the intricate world of scientific computing, ushering in a new era of rigor, reproducibility, and transparency. This blog post charts an uncharted course into the profound influence of DevOps on the scientific domain, highlighting the interplay between three pivotal pillars: Containers, Git Repositories, and CI/CD Engines. 🚀 Unveiling the Three Pillars 1. Containers: The Scientific Enigma Unraveled 🧩 Scientific computing has always been a puzzle of dependencies and configurations, where the need to replicate the perfect computational environment can be a Sisyphean task. Here, the enigma is containers, the all-encompassing key to unlock scientific progress. Containers encapsulate complete, reproducible computational environments that transcend physical limitations. This portability allows researchers to traverse platforms and infrastructures effortlessly, from local workstations to colossal computing clusters, even extending their reach to the celestial cloud, typified by Kubernetes. Containers, in their singularity, eliminate the complexities of setup, enabling a fluid exchange of knowledge and ideas among scientists. 📦 2. Git Repositories: The Alchemist's Chronicle 🔮 At the heart of scientific inquiry lies code – the recipe to unlock the secrets of data. This code represents the philosopher's stone, the core of every scientific investigation. It's imperative to preserve an unadulterated historical record of its transformation, a role diligently performed by Git repositories. Git, the venerable alchemist, offers distributed version control, proven invaluable in the realm of DevOps, and effortlessly adapted to scientific needs. These repositories, eternal chronicles of code evolution, foster transparency, collaboration, and dispute resolution. They are fortified with cryptographic timestamps, ensuring the immutability and legal sanctity of the code's history – a shield of authenticity in the uncertain world of scientific exploration. 🔗 3. CI/CD Engines: Forging a Seamless Frontier 🛠️ The heart of DevOps, the forge where software is crafted, is Continuous Integration (CI) and Continuous Delivery (CD). These engines optimize software development, yet their potential extends beyond the IT industry. CI involves the constant amalgamation of code changes, while CD automates deployment. These engines, like GitLab CI/CD, offer the architectural blueprint for scientific workflows. By orchestrating pipelines, CI/CD engines enable researchers to abstract away the intricacies of their environments. The orchestration and automation of these pipelines elevate scientific computing, making it a seamless venture. Researchers traverse the unknown frontier with ease, enabled by their CI/CD compass. 🚧 Integration Examples Three tales of integration illustrate the harmonious marriage of these three pillars in redefining scientific computing. 1. Centralized Pipelines: Navigating an Ocean of Consistency 🏢 Centralized pipelines, the cornerstone of computational science, unfold in the comfort of homogeneous environments – think high-performance computing clusters or cloud platforms. These pristine, managed sanctuaries harbor the magnificence of containerized DevOps workflows. Researchers, charting courses through these environments, find consistency, scalability, and elegance. The translation of scientific work across these platforms is no longer a Herculean task; it is a voyage of discovery in an ocean of computational consistency. 🌐 2. Decentralized Pipelines: Bridging Heterogeneous Horizons 🌍 For those who tread the uncharted path of scaling out an analysis, the DevOps torch illuminates their way. It's the bridge spanning the chasm between local exploration and the computational cluster. The faithful Docker container, partner-in-crime of scientists, carries the burden. It hosts their code, their dreams, and their ambitions, in an environment harmonized for both solitary exploration and grandiose cluster performances. DevOps equips researchers to quell scaling fears, test their might against localized adversaries, and seamlessly cross into the realm of computational giants. ⚖️ 3. Minimizing Scale Out Friction: A Journey of Fluid Transition 🧘 The journey from local testing grounds to the battlefield of scalability can be tumultuous. This transition is where shortcuts can lead to disaster, and where DevOps emerges as the guiding sage. The constant companion of researchers is the versatile Docker container, accommodating local forays, and large-scale skirmishes with equal ease. It allows the validation of pipelines in a controlled environment, a simulator for the grand clash on the computational cluster. DevOps ensures a transition devoid of strife, a harmonious crescendo to the symphony of scientific discovery. 📈 Beneficial Side Effects Beyond the immediate effects, the infusion of DevOps into scientific computing ripples with wider benefits. 1. Superior Rigor, Reproducibility, and Transparency 📚 DevOps tools elevate the standards of scientific rigor, reproducibility, and transparency. They provide a robust foundation for research, rendering scalability and parallelism seamless. The shared environments and Git repositories eliminate barriers to replicating experiments. This potent combination of DevOps tools and principles opens new horizons, ensuring that research is conducted with the utmost rigor and transparency. 📖 2. FAIR Principles and Interoperable APIs: A Symphony of Data Reuse 🔄 DevOps frameworks echo the harmonious notes of FAIR principles, heralding the era of interoperable research data. Through structured pipelines with rich metadata, datasets are primed for sharing and indexing, adhering to the principles of Findable, Accessible, Interoperable, and Reusable (FAIR) data. DevOps also paves the way for the creation of common, interoperable APIs across diverse scientific domains, accelerating collaboration and propelling scientific exploration. 🔄 3. Compliance With Open Data Mandates: A New Horizon 🌐 In an age where open data mandates are on the rise, DevOps tools ensure that researchers are well-prepared to meet these demands. The tools guide researchers on data release, simplifying the path to compliance. By seamlessly adopting DevOps practices, researchers align with open data mandates, enriching the scientific ecosystem with accessibility and transparency. 📊 Conclusion The confluence of DevOps and scientific computing is nothing short of a revolution. It reshapes the frontier of scientific inquiry, redefining how researchers navigate the intricacies of computational analysis. With DevOps for Scientific Computing, researchers can: Elevate scientific rigor, reproducibility, and transparency. Champion the adoption of FAIR principles and interoperable APIs. Embrace open data mandates with grace and ease. This revolution not only optimizes research workflows but also stands as a symbol of trust and integrity in a world where science's role in society is more crucial than ever. 🌍 In summary, is the transformative force that accelerates scientific progress, enhances the quality of research, and reinforces the integrity of scientific endeavors. DevOps for Scientific Computing It's the symphony where technology and knowledge harmoniously converge, ensuring the pursuit of scientific knowledge remains as captivating and enlightening as ever. 🌟 References Wolfgang Alschner. "The computational analysis of international law". Research Methods in International Law. Edward Elgar Publishing, 2021. Sumon Biswas, Mohammad Wardat, and Hridesh Rajan. "The art and practice of data science pipelines: A comprehensive study of data science pipelines in theory, in-the-small, and in-the-large". Proceedings of the 44th International Conference on Software Engineering, 2022. Jennifer Bryan. "Excuse me, do you have a moment to talk about version control?" The American Statistician, 2018. Yolanda Gil et al. "Examining the Challenges of Scientific Workflows". Computer, 2007. Gregory M Kurtzer, Vanessa Sochat, and Michael W Bauer. "Singularity: Scientific containers for mobility of compute". PLoS One, 2017. Christine Laine et al. "Reproducible research: moving toward research the public can really trust". Ann Intern Med, Mar. 2007. Karthik Ram. "Git can facilitate greater reproducibility and increased transparency in science". Source Code Biol Med, 2013. Roger D Peng. "Reproducible research in computational science". Science, Dec. 2011. Russell A Poldrack et al. "Scanning the horizon: towards transparent and reproducible neuroimaging research". Nat Rev Neurosci, Feb. 2017. Mark D Wilkinson et al. "The FAIR Guiding Principles for scientific data management and stewardship". Sci Data, Mar. 2016.

Transforming the Scientific Frontier: The DevOps Revolution

Fallout 4 has a big wasteland with plenty of things to kill. Between the Gunners and the Deathclaws, one can find plenty of targets to shoot at with bullets and lasers. The only problem is that there are only so many guns in the Commonwealth to shoot before you’ve shot them all. Fortunately, the Fallout 4 Nexus has several thousand mods to choose from. Of those several thousand, we have found nine Fallout 4 Weapons mods including some very new mods to give you a few new ways to kill your way to your son faster. Also, some of these weapon mods don’t really fit Fallout 4’s art style which won’t be a problem if you look at our list of . There’s something there to fit these new toys. armor mods Fallout 4 Weapon Mods List Classic Holstered Weapons System (CHW) Cazador Missile Launcher Glock 19x QMW (Quick Modification Weapon)(Only mod weapon) Cowboy Repeater AKA Winchester Model 1892 FACTOR - Modular Rifle DefenseGun Misriah Armory M1928A1 Thompson - Day of Infamy Classic Holstered Weapons System (CHW) by shavkacagarikia shavkacagarikia/Fallout 4 Nexus It’s a little odd to start off a weapons list without using one of the Fallout 4 Weapon Mods but this makes it all better. One of Fallout 4’s more disappointing aspects is the fact that weapons turn invisible when holstered, making it difficult to admire them when going from one part of the wasteland to another. In the early days before this mod, the solution was to equip the weapons as apparel, allowing them to be visible in a janky, unimmersive way. Shavkacagarikia has gotten around this through some form of black magic, introducing a system that properly allows weapons to appear when holstered. It even works on modded weapons! If the way the weapon is holstered looks too janky, there is an in-game tool to tweak its positioning to make sure that it looks just right. Download link: Fallout 4 Nexus Cazador Missile Launcher by TheBigLidowski Utherien/Fallout 4 Nexus The Cazador, modeled after the real-world MATADOR recoilless rifle, is essentially a new kind of missile launcher. The weapon brings some modern flavour to Fallout with its own animations, sounds and integration with the game’s spawns. The Cazador is highly customisable with 13 different ammunition types that can be swapped in the field, giving the weapon plenty of versatility especially considering the unique damage types. It also comes with four scopes with their own functions and eight cosmetic skins. Download Link: Fallout 4 Nexus Glock 19x by Ardent Ardent117/Fallout 4 Nexus You might have a fondness for Wasteland’s selection of American pistols but I can recommend Ardent117’s Austrian varieties. The Glock 19X’s recency easily makes it one of the more modern Fallout 4 Weapon Mods. The weapon benefits from all of the modern world’s weapon customisation. Just about any part of the weapon can be customised with meaningful gameplay changes. The custom sounds and animations really sell the weapon and integrates it perfectly with the game. Q.M.W (Quick Modification Weapon)(Only mod weapon) by Tooun Tooun/Fallout 4 Nexus This mod doesn’t add any weapons but it lets you modify your weapons on the go, if the weapon is compatible with it. The compatible weapons tend to be modern, high quality mods that can be modified to suit many situations. Using this mod makes it possible to take full advantage of the modifications, letting the weapons get more versatile and adds some variety to gameplay. In practice, it is quite similar to weapon customisation in Crysis 3 and Metro Exodus. The menu’s interface doesn’t look remotely lore friendly but it more than makes up for it in functionality. Download Link: Fallout 4 Nexus Cowboy Repeater AKA Winchester Model 1892 by lucasncolombo lucasncolombo/Fallout 4 Nexus Lever actions are rare in the world of Fallout 4 Weapon Mods which makes this mod a treat. It features fantastic textures that bring out a rugged look that is fit for the wasteland. It’s got the custom animations to match its style. The weapon can’t be acquired easily. Instead, it has to be found in its ruined state and repaired after finding a crafting book. Download Link: Fallout 4 Nexus FACTOR - Modular Rifle by VigVam33 VigVam33/Fallout 4 Nexus The Modular Rifle truly lives up to its name. This weapon can be modified to the extent that it can be changed from a laser weapon to a ballistic weapon. It’s got variable zoom sights, multiple barrel lengths with 200 weapon mods. This mod pretty much lets you make a weapon perfect for your needs. DefenseGun by Yona YonaTaku/Fallout 4 Nexus DefenceGun is a fairly old mod that still stands up to its newer counterparts. It adds a weapon platform that resembles the old British Sten from the Second World War. The side-loaded magazine adds a unique touch to the weapon. It’s also available from lower levels, letting it give some variety and a break from the pipe weapons. Misriah Armory NovaFinch/Nexus Misriah Armory is one of the best Fallout 4 Weapon Mods that draw on fictional weapons. The quality of its textures and customisablity are some of the best on the Nexus. We also took a much deeper look at this mod before. Download Link: Fallout 4 Nexus M1928A1 Thompson - Day of Infamy by NWI - Noniac - Wardaddy - Ricoginger Noniac/Fallout 4 Nexus Fallout 4 technically has this weapon in the form of the Submachine Gun. Unfortunately, it doesn’t look half as good and doesn’t have the same feel of this classic weapon. The textures themselves are taken from Day of Infamy which look amazing. You can find it on Triggermen from level 12 and Bosses from Level 20 Final thoughts on these Best Fallout 4 Weapon Mods These nine Fallout 4 Weapon Mods are some of the best mods to come out, especially considering the amazing selection from the last year. These mods will spice up your game with some new toys that will give some life to the game. Also, be sure to check out: Warhammer Age of Sigmar: Storm Ground Arrives May 27, New Trailer Out The 25 Funniest Roblox Memes in 2021 Final Fantasy XIV Reshade Guide

9 Best Fallout 4 Weapons Mods in 2021

The story begins on a cold winter morning in New York where a small group of people are meeting in a windowless room in an unremarkable corner of UN headquarters. It is the job of these people to communicate humanitarian issues, and in a world on fire, they have their work cut out for them. They have gathered to watch a film called about a Syrian refugee camp. It is February 2014. Virtual Reality Clouds Over Sidra Like everyone else in the room I have heard a lot about Virtual Reality (VR), but this is the first time I have watched a VR film. I am a journalist turned filmmaker and have been telling humanitarian stories in words, pictures and film for 20 years. I have worked and filmed in the Za’atari refugee camp before and have made, or helped to make, several films about the plight of Syrian refugees. I know what to expect. Or so I think. Six minutes later I lay the headset back on the table and take a deep breath. I’m stunned by the way Gabo Arora’s film has dragged me into the heart of his story and I’m daring to think that this could be the storytelling that I’ve been waiting for. It feels like both the end and beginning of a journey. It feels like a glimpse of the future. tool Fast forward 18 months and I’m sitting in an airless refugee tent in Lebanon’s Bekaa Valley. On one side of me sits Nokia’s beautiful Ozo VR camera and my long term cameraman and friend Thomas, and on the other side, former Iraqi refugee-turned-aid-worker Warda and half a dozen young Syrian refugee girls. We are passing round a Samsung Gear headset loaded with a short animated film about dinosaurs to try and explain VR to them. Click play to see what happened when they watched VR for the first time … A long time before I’d ever heard of VR I read somewhere that “the evolution of storytelling is inseparable from the evolution of technology.” I made a note of it at the time and promptly forgot about it, but it was a thought that kept coming back to me during the shoot of our VR film, . Life in the Time of Refuge I had been warned by friends who had been in the VR game for a while already that shooting in VR meant throwing out pretty much everything you think you know about filmmaking and starting over. “ You have to hide to be out of shot … You can’t follow the action … You can’t frame the shot because everything is in shot … And whatever you do, be very careful of movement, because even the slightest movement can make the viewer feel sea sick …” So when Thomas suggested one morning while we were filming by a lake in Finland that we set up the camera in a boat and “get a shot in the middle of the lake,” that sounded like trouble to me. Click play to see what happened when we tried … For sure, it’s way more difficult and time consuming to shoot documentary stories in VR, and the run ’n’ gun style of observational documentary is nigh on impossible (at least for now). Not to mention the challenges of shooting in areas where insecurity is high and people are suspicious of strange people, doing strange things with strange looking bits of kit. Click play below to see what it was like shooting a VR scene close to the Lebanon/Syria border … A s it turns out — and, as I had been warned — much of shooting VR is about jettisoning the things you love about filmmaking and embracing the things you thought you didn’t like. Things like voice over. Remember Robert McKee’s words of warning to a paranoid Nicholas Cage in Charlie Kaufmann’s ?: Adaptation “God help you if you use voice-over in your work, my friends. God help you! That’s flaccid, sloppy writing. Any idiot can write voice-over narration to explain the thoughts of the character.” I thought that too many of the humanitarian VR films that I had seen to date were too reliant on voice over, and that the story needed a chance to develop alongside its characters. With my snooty filmmaker hat on, I saw voice over as the last refuge of the scoundrel so when we started the project I was adamant that we should avoid it. But as we found out, the bottom line is that in VR so much of the director’s ability to let the narrative take its own course is taken out of the equation by the fact that the director and cinematographer are hiding behind a grave at the crucial moment. So in the end we bit the bullet and embraced some ‘flaccid, sloppy writing.’ But you know what, we think it worked out pretty well … Click play to watch the 360 version of our film on Facebook … ends

Shooting VR or: How I learned to stop worrying and love the voice over

Social Media and the Battlefield: Human Error in Modern Warfare

Analyzing Audience Responses to Ukrainian Memetic Warfare Tactics

Authors:
(1) Yelena Mejova, ISI Foundation, Italy;
(2) Arthur Capozzi, Università degli Studi di Torino, Italy;
(3) Corrado Monti, CENTAI, Italy;
(4) Gianmarco De Francisci Morales, CENTAI, Italy. Table of Links Introduction Background and Related Work Methods Results Discussion and References 3 METHODS 3.1 Data collection To examine the use of memetic engagement by Ukraine during the first 8 months of the 2022 conflict, we begin with a large collection of tweets matching war-related keywords (see Appendix A) spanning from 27 February 2022 to 12 October 2022. By examining this collection, we identify three prominent accounts: (1) @Ukraine: self-described as “the official Twitter account of Ukraine”, this verified account is labeled as an “Ukraine government organization” by Twitter. It posts war-related and political updates about Ukraine. (2) @uamemesforces: self-described as the “Source of the best Ukrainian memes”, this unverified account states as its goal ”to convey the truth about what is happening now in Ukraine with the help of memes”.[8] (3) @DefenceU: or “Defense of Ukraine”, is the “Official page of the Ministry of Defense of Ukraine”. Like @Ukraine, this verified account is described by Twitter as a “Ukraine government organization”. Its posts concern battlefield updates and news. We select the accounts @Ukraine and @DefenceU because of their large following (at the end of the collection period), because they are official accounts not associated with individuals or specific media outlets, and because of the focus on the conflict (many popular Ukrainian accounts are by non-relevant celebrities). Additionally, we select @uamemesforces, the most popular Ukrainian unofficial account that posts memes about the conflict. Note that we are limited in the number of accounts we can study due to the labor-intensive labeling process. We retrieve all of the posts for these accounts via Twitter Timeline API. We treat each post with media content as a potential meme, no matter the format. To ensure preserving content posted by these accounts, we also download the images associated with the posts and the users who retweeted or liked the post (see Table 1 for summary statistics). We associate each user with a country in the GeoNames database via their location field (a free-text field optionally filled by the user). Considering only tweets with media, out of 586 042 users, 44% are successfully mapped to a country. We retain only countries with at least 1000 geolocated retweeting users for geographic analysis. In order to explore the relationship between Ukrainian memes and each country’s actions towards the conflict, we use the Ukraine Support Tracker data from the Kiel Institute for the World Economy (downloaded on 20 June 2023), which “quantifies military, financial, and humanitarian aid transferred by governments to Ukraine since the end of diplomatic relations between Russia and Ukraine on January 24, 2022”.[9] The dataset provides a quantitative measure of a country’s aid to Ukraine. Finally, to examine the attitude of each country’s population towards Ukraine, we use the Eurobarometer STD98 survey which took place in winter 2022-2023 in 30 European countries [22]. Specifically, we use use QE2, answer 6: “To what extent do you agree or disagree with: providing financial support to Ukraine.” 3.2 Annotation To understand the visual and narrative attributes of the captured data, all authors perform a manual annotation of a random selection of the tweets. We use all of @Ukraine tweets and a random sample of roughly 40% of those from the other two accounts are selected for analysis, for a total of 1063 tweets (last column of Table 1). We remove self-replies (an account replying to its own post) which have no retweets of their own. Guided by previous literature, we compose a codebook to capture visual and narrative aspects of the content. First, we adopt the significant features of the content characterization framework proposed by Ling et al. [41], which is used to model the virality of political memes. These attributes include: (1) Number of panels: multiple or single; (2) Type of image: photo, screenshot, or illustration; (3) Scale: close-up, medium shot, long shot, other; (4) Type of subject: object, character, scene, creature, text, other; (5) Main attribute of the subject (if character): facial expression, posture, poster (6) Character emotion (if character): positive, negative, or neutral; (7) Contains words: whether image itself (not tweet caption) contains words. Second, we employ the narrative framework by Karpman [35] later extended by de Saint Laurent et al. [18]. It defines two key dimensions to characterize a narrative: moral quality (benevolent or malevolent) and power (strong or weak), thus resulting in four character archetypes: heroes (benevolent, strong), victims (benevolent, weak), villains (malevolent, strong), and fools (malevolent, weak). We also label the actors mentioned in these narratives (as an open category, guided by an initial set of main actors in the conflict, such as Zelensky and Putin). We further contextualize the narratives by their emotional appeal [8], in particular humor, pride, fear, outrage, and compassion (the label was originally open, but the labelers coalesced on this set). Finally, we record whether there is a specific intent in terms of a call to action or information sharing. Note that, as much as possible, we try to label the narratives as intended from the point of view of the posting account (although some biases may creep in, as we discuss in Section 5). Appendix B reports the full codebook. Using this literature-driven codebook, all authors of this manuscript perform a deductive coding of the selected posts (for more on such process see Linneberg and Korsgaard [42]). We rely on Google Translate to translate non-English language content. All coders first annotated a selection of 50 posts, discussed insights and disagreements, and came up with standard labels for emotional appeal, intent, and actor (though the label remained open). Subsequently, the rest of the posts are annotated by one author each, with uncertain cases discussed collectively. Appendix C reports the inter-annotator agreement on a selection of 48 posts (sampled from each account). On average, the agreement is 0.750 for @Ukraine, 0.604 for @uamemesforces, and 0.667 for @DefenceU (with a Krippendorff’s alpha of around 0.65 for all accounts and tasks), which is high considering the subjective nature of the tasks, and that some features were open and could have multiple labels. The annotated dataset is available at https://github.com/corradomonti/ukranian-tweets. 3.3 Popularity model Using the data collected and annotated as per the previous sections, we construct models to find whether there are trends in (𝑖) what kind of content becomes more popular, and (𝑖𝑖) who it reaches. Specifically, we use linear regression to model the log-transformed number of retweets with the content variables outlined above as independent variables. As the three accounts have posted a vastly different number of tweets, we weigh the data points (tweets) inverseproportionally to their representation in the dataset, so that each account weighs equally on the outcomes. We build several versions of the model to ascertain the benefit of each class of variables: visual [41], actor, emotion, intent, and narrative. A baseline model simply uses the identity of the author (i.e., group fixed effects), the number of followers of that account (log transformed) as a proxy for the potential audience, and the number of days from the invasion to account for changing interest in the topic over time. We then choose the best model by the Bayesian Information Criterion (BIC)—a model selection metric that captures both unexplained variation in the dependent variable and the number of explanatory variables. The coefficients of this model reveal the associations between these variables and the virality of the content. Next, we group retweets by the country of the user (those who can be geolocated) to understand the preferences of users who spread these messages. We enrich our characterization of the relationships between each country and Ukraine by considering the amount of financial assistance the country provides to Ukraine (normalized by the country’s GDP) or by their opinion on this topic (via the Eurobarometer survey). As mentioned above, we consider these measures as different proxies of support for Ukraine’s military efforts. First, we examine the retweeting rate (normalized by population). Then, we compare the narrative preferences to the amount of support for Ukraine: are there narratives associated with increased support for Ukraine? We operationalize this preference as the log odds of users from a country retweeting content having a specific narrative. Finally, we examine the geographical distribution of the retweeting rate and narrative preferences across Europe by plotting these on choropleth maps. [8] https://www.patreon.com/uamemesforces [9] https://www.ifw-kiel.de/topics/war-against-ukraine/ukraine-support-tracker This paper is available on arxiv under CC 4.0 license. Authors: (1) Yelena Mejova, ISI Foundation, Italy; (2) Arthur Capozzi, Università degli Studi di Torino, Italy; (3) Corrado Monti, CENTAI, Italy; (4) Gianmarco De Francisci Morales, CENTAI, Italy. Authors: (1) Yelena Mejova, ISI Foundation, Italy; (2) Arthur Capozzi, Università degli Studi di Torino, Italy; (3) Corrado Monti, CENTAI, Italy; (4) Gianmarco De Francisci Morales, CENTAI, Italy. Table of Links Introduction Introduction Background and Related Work Background and Related Work Methods Methods Results Results Discussion and References Discussion and References 3 METHODS 3.1 Data collection To examine the use of memetic engagement by Ukraine during the first 8 months of the 2022 conflict, we begin with a large collection of tweets matching war-related keywords (see Appendix A) spanning from 27 February 2022 to 12 October 2022. By examining this collection, we identify three prominent accounts: (1) @Ukraine: self-described as “the official Twitter account of Ukraine”, this verified account is labeled as an “Ukraine government organization” by Twitter. It posts war-related and political updates about Ukraine. @Ukraine: (2) @uamemesforces: self-described as the “Source of the best Ukrainian memes”, this unverified account states as its goal ”to convey the truth about what is happening now in Ukraine with the help of memes”.[8] @uamemesforces: (3) @DefenceU: or “Defense of Ukraine”, is the “Official page of the Ministry of Defense of Ukraine”. Like @Ukraine, this verified account is described by Twitter as a “Ukraine government organization”. Its posts concern battlefield updates and news. @DefenceU: We select the accounts @Ukraine and @DefenceU because of their large following (at the end of the collection period), because they are official accounts not associated with individuals or specific media outlets, and because of the focus on the conflict (many popular Ukrainian accounts are by non-relevant celebrities). Additionally, we select @uamemesforces, the most popular Ukrainian unofficial account that posts memes about the conflict. Note that we are limited in the number of accounts we can study due to the labor-intensive labeling process. We retrieve all of the posts for these accounts via Twitter Timeline API. We treat each post with media content as a potential meme, no matter the format. To ensure preserving content posted by these accounts, we also download the images associated with the posts and the users who retweeted or liked the post (see Table 1 for summary statistics). We associate each user with a country in the GeoNames database via their location field (a free-text field optionally filled by the user). Considering only tweets with media, out of 586 042 users, 44% are successfully mapped to a country. We retain only countries with at least 1000 geolocated retweeting users for geographic analysis. In order to explore the relationship between Ukrainian memes and each country’s actions towards the conflict, we use the Ukraine Support Tracker data from the Kiel Institute for the World Economy (downloaded on 20 June 2023), which “quantifies military, financial, and humanitarian aid transferred by governments to Ukraine since the end of diplomatic relations between Russia and Ukraine on January 24, 2022”.[9] The dataset provides a quantitative measure of a country’s aid to Ukraine. Finally, to examine the attitude of each country’s population towards Ukraine, we use the Eurobarometer STD98 survey which took place in winter 2022-2023 in 30 European countries [22]. Specifically, we use use QE2, answer 6: “To what extent do you agree or disagree with: providing financial support to Ukraine.” 3.2 Annotation To understand the visual and narrative attributes of the captured data, all authors perform a manual annotation of a random selection of the tweets. We use all of @Ukraine tweets and a random sample of roughly 40% of those from the other two accounts are selected for analysis, for a total of 1063 tweets (last column of Table 1). We remove self-replies (an account replying to its own post) which have no retweets of their own. Guided by previous literature, we compose a codebook to capture visual and narrative aspects of the content. First, we adopt the significant features of the content characterization framework proposed by Ling et al. [41], which is used to model the virality of political memes. These attributes include: (1) Number of panels: multiple or single; (2) Type of image: photo, screenshot, or illustration; (3) Scale: close-up, medium shot, long shot, other; (4) Type of subject: object, character, scene, creature, text, other; (5) Main attribute of the subject (if character): facial expression, posture, poster (6) Character emotion (if character): positive, negative, or neutral; (7) Contains words: whether image itself (not tweet caption) contains words. Second, we employ the narrative framework by Karpman [35] later extended by de Saint Laurent et al. [18]. It defines two key dimensions to characterize a narrative: moral quality (benevolent or malevolent) and power (strong or weak), thus resulting in four character archetypes: heroes (benevolent, strong), victims (benevolent, weak), villains (malevolent, strong), and fools (malevolent, weak). We also label the actors mentioned in these narratives (as an open category, guided by an initial set of main actors in the conflict, such as Zelensky and Putin). We further contextualize the narratives by their emotional appeal [8], in particular humor, pride, fear, outrage, and compassion (the label was originally open, but the labelers coalesced on this set). Finally, we record whether there is a specific intent in terms of a call to action or information sharing. Note that, as much as possible, we try to label the narratives as intended from the point of view of the posting account (although some biases may creep in, as we discuss in Section 5). Appendix B reports the full codebook. Using this literature-driven codebook, all authors of this manuscript perform a deductive coding of the selected posts (for more on such process see Linneberg and Korsgaard [42]). We rely on Google Translate to translate non-English language content. All coders first annotated a selection of 50 posts, discussed insights and disagreements, and came up with standard labels for emotional appeal, intent, and actor (though the label remained open). Subsequently, the rest of the posts are annotated by one author each, with uncertain cases discussed collectively. Appendix C reports the inter-annotator agreement on a selection of 48 posts (sampled from each account). On average, the agreement is 0.750 for @Ukraine, 0.604 for @uamemesforces, and 0.667 for @DefenceU (with a Krippendorff’s alpha of around 0.65 for all accounts and tasks), which is high considering the subjective nature of the tasks, and that some features were open and could have multiple labels. The annotated dataset is available at https://github.com/corradomonti/ukranian-tweets. 3.3 Popularity model Using the data collected and annotated as per the previous sections, we construct models to find whether there are trends in (𝑖) what kind of content becomes more popular, and (𝑖𝑖) who it reaches. Specifically, we use linear regression to model the log-transformed number of retweets with the content variables outlined above as independent variables. As the three accounts have posted a vastly different number of tweets, we weigh the data points (tweets) inverseproportionally to their representation in the dataset, so that each account weighs equally on the outcomes. We build several versions of the model to ascertain the benefit of each class of variables: visual [41], actor, emotion, intent, and narrative. A baseline model simply uses the identity of the author (i.e., group fixed effects), the number of followers of that account (log transformed) as a proxy for the potential audience, and the number of days from the invasion to account for changing interest in the topic over time. We then choose the best model by the Bayesian Information Criterion (BIC)—a model selection metric that captures both unexplained variation in the dependent variable and the number of explanatory variables. The coefficients of this model reveal the associations between these variables and the virality of the content. Next, we group retweets by the country of the user (those who can be geolocated) to understand the preferences of users who spread these messages. We enrich our characterization of the relationships between each country and Ukraine by considering the amount of financial assistance the country provides to Ukraine (normalized by the country’s GDP) or by their opinion on this topic (via the Eurobarometer survey). As mentioned above, we consider these measures as different proxies of support for Ukraine’s military efforts. First, we examine the retweeting rate (normalized by population). Then, we compare the narrative preferences to the amount of support for Ukraine: are there narratives associated with increased support for Ukraine? We operationalize this preference as the log odds of users from a country retweeting content having a specific narrative. Finally, we examine the geographical distribution of the retweeting rate and narrative preferences across Europe by plotting these on choropleth maps. [8] https://www.patreon.com/uamemesforces [9] https://www.ifw-kiel.de/topics/war-against-ukraine/ukraine-support-tracker This paper is available on arxiv under CC 4.0 license. This paper is available on arxiv under CC 4.0 license. available on arxiv

Understanding Memetic Warfare Tactics: Social Media Tactics in the Russo-Ukrainian War

Authors:
(1) Yelena Mejova, ISI Foundation, Italy;
(2) Arthur Capozzi, Università degli Studi di Torino, Italy;
(3) Corrado Monti, CENTAI, Italy;
(4) Gianmarco De Francisci Morales, CENTAI, Italy. Table of Links Abstract & Introduction Related Work Methodology Experiment Result Discussion and References The 2022 Russian invasion of Ukraine has seen an intensification in the use of social media by governmental actors in cyber warfare. Wartime communication via memes has been a successful strategy used not only by independent accounts such as @uamemesforces, but also—for the first time in a full-scale interstate war—by official Ukrainian government accounts such as @Ukraine and @DefenceU. We study this prominent example of memetic warfare through the lens of its narratives, and find them to be a key component of success: tweets with a ‘victim’ narrative garner twice as many retweets. However, malevolent narratives focusing on the enemy resonate more than those about heroism or victims with countries providing more assistance to Ukraine. Our findings present a nuanced examination of Ukraine’s influence operations and of the worldwide response to it, thus contributing new insights into the evolution of socio-technical systems in times of war. 1 INTRODUCTION On 24 February 2022, Russia began a large-scale military invasion of Ukraine. The United Nations Human Rights Office has confirmed the death of 9083 civilians up to 18 June 2023.[1] However, their report highlights that the actual figures are believed to be considerably higher, and most estimates suggest more than 300 000 total deaths. In the midst of this immense tragedy, the two governments at war—Russia and Ukraine—are also waging an information war on many fronts. Due to the prevalence of the Internet as a primary information source for Ukrainians,[2] the war has been described as the first social-media war,[3] echoing the reputation of the Vietnam war as the “first television war” due to the role this medium had in informing the public about the conflict. Both governments and common people have been using social media to propagate their views on the conflict and affect global public opinion. A vivid example is @Ukraine, the official account of the Ukrainian government on Twitter.[4] This account was started in 2016 by Yarema Dukh while working as a press officer in the Ukrainian presidential office. One of the memes produced by this account, depicting “living next to Russia” as the worst possible headache,[5] has been shared more than 600 000 times and has reached 55 million views. According to the author, the goal of this meme was “to explain to some large and distant target audiences that Russia is the problem here, not Ukraine, the West, the U.S., NATO, aliens or anyone else”.[6] This account has posted a variety of content and has now become one of the most visible examples of memetic warfare from a warring government. An early definition of memetic warfare was given by the NATO Strategic Communications Centre of Excellence in Latvia in a document titled “It’s time to embrace memetic warfare” [25]. The document defines it as “the competition over narrative, ideas, and social control in a social-media battlefield” and is viewed as “the ’digital native’s version of psychological warfare, more commonly known as propaganda”. The authors argue that it needs to be brought into mainstream military thinking by investing in its development. While propaganda has long been a central component of war, memetic warfare introduces a new key aspect: its reach is intrinsically tied to how much its own audience is willing to spread it. Governments involved in psychological warfare used to rely purely on one-way communication media such as posters and radio. However, the two-way nature of social media implies that they have to convince their own audience to help spread their message to reach wider audiences. This characteristic creates a particular amalgam of social-media bottom-up dynamics with the top-down approach typical of military objectives and narratives. Different goals then might require spreading different types of content, which in turn might have varying degrees of success with specific audiences. Despite the importance of this matter, so far little research effort has been dedicated to the quantitative analysis of which aspects make memetic warfare more successful, which audience it is able to reach, and what are its consequences. These questions are of primary interest not only for the involved parties, but also for the global public opinion. In order to prepare, evaluate, and react appropriately, we need to better understand these novel techniques of propaganda, which are likely to become increasingly common in the future. Our analysis focuses on these research questions: RQ1 What are the main characteristics of memes shared by popular Ukrainian Twitter accounts? RQ2 Which attributes of a wartime meme are good predictors of its virality? RQ3 What is the geographic resonance of each narrative, and how does it relate to each country’s actions toward the conflict? We examine the content of the memes in terms of the actors involved, their emotional appeal, and the narratives used. Based on previous literature [18], we identify four roles within a narrative, depending on the moral quality of the actors in the narrative and on their perceived power. In our context, these roles often represent the benevolent moral quality of the author’s side or the malevolent one of the enemy. The narratives we identify are a powerful predictor of the virality of a meme. Content with a victim narrative—a benevolent, weak actor—has a 109% higher reach w.r.t. content with no narrative, while a villain narrative—depicting the enemy as malevolent and strong—increases reach by 32%. Finally, since one of the goals of these accounts is to influence global public opinion, we look at the success of the content geographically. Memetic warfare reach reflects important aspects of the offline world: the retweets per inhabitant in a given country correlate highly (𝜌 = 0.787) with the assistance given by that country to Ukraine. In addition, different narratives resonate differently with different countries. For instance, villain narratives resonate more than victim narratives in European countries bordering Russia. In turn, these different framings of the conflict are also reflected in the material assistance given by these countries. The success of malevolent narratives positively correlates with the share of GDP devoted to assisting Ukraine (𝜌 = 0.601, 𝜌 = 0.566). The present study has important implications for our understanding of the interaction between communication technologies and their use by governments and non-governmental actors during a conflict. We illustrate how social media may factor into a country’s strategic messaging during war, and how such messaging is related to the attention and aid afforded to the country. As cyber-influence campaigns become more common and sophisticated, we encourage the research community to apply a multi-factorial view of the relevant content, including visual and narrative aspects, to understand its resonance around the world. [1] https://www.ohchr.org/en/news/2023/06/ukraine-civilian-casualty-update-19-june-2023 [2] https://www.unian.info/society/media-space-internet-bypasses-tv-as-main-news-source-for ukrainians-11330729.html [3] https://www.forbes.com/sites/petersuciu/2022/03/01/is-russias-invasion-of-ukraine-the-first social-media-war [4] Twitter is currently being rebranded to “X”. For consistency with previous literature, we prefer using the former name. [5] https://twitter.com/Ukraine/status/1468206078940823554 [6] https://www.washingtonpost.com/world/2022/01/26/ukraine-russia-memes-social This paper is available on arxiv under CC 4.0 license. Authors: (1) Yelena Mejova, ISI Foundation, Italy; (2) Arthur Capozzi, Università degli Studi di Torino, Italy; (3) Corrado Monti, CENTAI, Italy; (4) Gianmarco De Francisci Morales, CENTAI, Italy. Authors: (1) Yelena Mejova, ISI Foundation, Italy; (2) Arthur Capozzi, Università degli Studi di Torino, Italy; (3) Corrado Monti, CENTAI, Italy; (4) Gianmarco De Francisci Morales, CENTAI, Italy. Table of Links Abstract & Introduction Abstract & Introduction Related Work Related Work Methodology Methodology Experiment Result Experiment Result Discussion and References Discussion and References The 2022 Russian invasion of Ukraine has seen an intensification in the use of social media by governmental actors in cyber warfare. Wartime communication via memes has been a successful strategy used not only by independent accounts such as @uamemesforces, but also—for the first time in a full-scale interstate war—by official Ukrainian government accounts such as @Ukraine and @DefenceU. We study this prominent example of memetic warfare through the lens of its narratives, and find them to be a key component of success: tweets with a ‘victim’ narrative garner twice as many retweets. However, malevolent narratives focusing on the enemy resonate more than those about heroism or victims with countries providing more assistance to Ukraine. Our findings present a nuanced examination of Ukraine’s influence operations and of the worldwide response to it, thus contributing new insights into the evolution of socio-technical systems in times of war. 1 INTRODUCTION On 24 February 2022, Russia began a large-scale military invasion of Ukraine. The United Nations Human Rights Office has confirmed the death of 9083 civilians up to 18 June 2023.[1] However, their report highlights that the actual figures are believed to be considerably higher, and most estimates suggest more than 300 000 total deaths. In the midst of this immense tragedy, the two governments at war—Russia and Ukraine—are also waging an information war on many fronts. Due to the prevalence of the Internet as a primary information source for Ukrainians,[2] the war has been described as the first social-media war,[3] echoing the reputation of the Vietnam war as the “first television war” due to the role this medium had in informing the public about the conflict. Both governments and common people have been using social media to propagate their views on the conflict and affect global public opinion. A vivid example is @Ukraine, the official account of the Ukrainian government on Twitter.[4] This account was started in 2016 by Yarema Dukh while working as a press officer in the Ukrainian presidential office. One of the memes produced by this account, depicting “living next to Russia” as the worst possible headache,[5] has been shared more than 600 000 times and has reached 55 million views. According to the author, the goal of this meme was “to explain to some large and distant target audiences that Russia is the problem here, not Ukraine, the West, the U.S., NATO, aliens or anyone else”.[6] This account has posted a variety of content and has now become one of the most visible examples of memetic warfare from a warring government. An early definition of memetic warfare was given by the NATO Strategic Communications Centre of Excellence in Latvia in a document titled “It’s time to embrace memetic warfare” [25]. The document defines it as “the competition over narrative, ideas, and social control in a social-media battlefield” and is viewed as “the ’digital native’s version of psychological warfare, more commonly known as propaganda”. The authors argue that it needs to be brought into mainstream military thinking by investing in its development. While propaganda has long been a central component of war, memetic warfare introduces a new key aspect: its reach is intrinsically tied to how much its own audience is willing to spread it. Governments involved in psychological warfare used to rely purely on one-way communication media such as posters and radio. However, the two-way nature of social media implies that they have to convince their own audience to help spread their message to reach wider audiences. This characteristic creates a particular amalgam of social-media bottom-up dynamics with the top-down approach typical of military objectives and narratives. Different goals then might require spreading different types of content, which in turn might have varying degrees of success with specific audiences. Despite the importance of this matter, so far little research effort has been dedicated to the quantitative analysis of which aspects make memetic warfare more successful, which audience it is able to reach, and what are its consequences. These questions are of primary interest not only for the involved parties, but also for the global public opinion. In order to prepare, evaluate, and react appropriately, we need to better understand these novel techniques of propaganda, which are likely to become increasingly common in the future. Our analysis focuses on these research questions: RQ1 What are the main characteristics of memes shared by popular Ukrainian Twitter accounts? RQ2 Which attributes of a wartime meme are good predictors of its virality? RQ3 What is the geographic resonance of each narrative, and how does it relate to each country’s actions toward the conflict? We examine the content of the memes in terms of the actors involved, their emotional appeal, and the narratives used. Based on previous literature [18], we identify four roles within a narrative, depending on the moral quality of the actors in the narrative and on their perceived power. In our context, these roles often represent the benevolent moral quality of the author’s side or the malevolent one of the enemy. The narratives we identify are a powerful predictor of the virality of a meme. Content with a victim narrative—a benevolent, weak actor—has a 109% higher reach w.r.t. content with no narrative, while a villain narrative—depicting the enemy as malevolent and strong—increases reach by 32%. Finally, since one of the goals of these accounts is to influence global public opinion, we look at the success of the content geographically. Memetic warfare reach reflects important aspects of the offline world: the retweets per inhabitant in a given country correlate highly (𝜌 = 0.787) with the assistance given by that country to Ukraine. In addition, different narratives resonate differently with different countries. For instance, villain narratives resonate more than victim narratives in European countries bordering Russia. In turn, these different framings of the conflict are also reflected in the material assistance given by these countries. The success of malevolent narratives positively correlates with the share of GDP devoted to assisting Ukraine (𝜌 = 0.601, 𝜌 = 0.566). The present study has important implications for our understanding of the interaction between communication technologies and their use by governments and non-governmental actors during a conflict. We illustrate how social media may factor into a country’s strategic messaging during war, and how such messaging is related to the attention and aid afforded to the country. As cyber-influence campaigns become more common and sophisticated, we encourage the research community to apply a multi-factorial view of the relevant content, including visual and narrative aspects, to understand its resonance around the world. [1] https://www.ohchr.org/en/news/2023/06/ukraine-civilian-casualty-update-19-june-2023 [2] https://www.unian.info/society/media-space-internet-bypasses-tv-as-main-news-source-for ukrainians-11330729.html [3] https://www.forbes.com/sites/petersuciu/2022/03/01/is-russias-invasion-of-ukraine-the-first social-media-war [4] Twitter is currently being rebranded to “X”. For consistency with previous literature, we prefer using the former name. [5] https://twitter.com/Ukraine/status/1468206078940823554 [6] https://www.washingtonpost.com/world/2022/01/26/ukraine-russia-memes-social This paper is available on arxiv under CC 4.0 license. This paper is available on arxiv under CC 4.0 license. available on arxiv

Narratives of War: Dissecting Ukrainian Memetic Warfare on Twitter

Q: What do a revolutionary from the 1800s, a supercomputer from the 1980s, and a tech start-up which supports decentralisation have in common? A: They’re all fictional. 🤣 Hilarious, right? Introducing the Luddites Let’s tie these things together. First of all, the revolutionary above refers to the folkloric figure of Ned Ludd, also known variously as General, Captain, and King Ludd. Whether there was ever a real Ned Ludd is very much up for debate. Sadly, since the Luddites were pretty much wiped out a couple of centuries ago by the combination of sending the military after them and executing them when captured (they threatened the almighty profit margin after all), we’re not likely to get an answer any time soon. What’s important for this article is that Ned Ludd was an idea. He was a mythical figurehead of a decentralised movement working to protect the livelihood of the individual worker, safeguarding their autonomy in the face of capitalist mechanical automation which worked to reduce them (or rather their cheaper, unskilled replacements) to cogs in a commercial machine. The Luddites are far from the first decentralised movement, and definitely not the last. They were unique, to my knowledge, in being the first to assemble a decentralised movement to oppose centralisation and concentration of power in the hands of a few. Those few being the capitalists who purchased the machines with the aim of exploiting their workforce by removing the need for expensive expertise. The Luddites were presented then, and are again now, as straightforward anti-progress revolutionaries. Even now the name is used as an insult, suggesting someone obstructs progress for the sake of obstruction. Ironically, those at the cutting edge of technologies designed to protect individual autonomy and identity have more in common with the Luddites than they do with most of the high-tech start-ups who claim to be fighting their corners. And the Supercomputer? In 1983 a film came out which became a cornerstone of hacker culture. No, not , I’m talking about the rightly-famous . Aside from being an innovative film, provides a brilliant illustration both of the danger of centralising power and authority, and of placing too much trust in technological infallibility. Hackers Wargames Wargames At the centre of the film is a supercomputer, Joshua (or more formally the WOPR), which controls the entire US nuclear arsenal and defensive systems. Even if you haven’t seen it, you can probably predict the rest of the film just from that. The centralisation piece is particularly striking when you know three names which deserve to go down in history (and likely others we don’t know): Stanislav Petrov, Vasily Arkhipov, and Sir David Ormsby-Gore. Their stories are worth looking up, but for this article what’s important to know is that it was their individual decision-making, their exercise of their autonomy, that means we’re not living in a nuclear hellscape today. As Joshua famously writes in , when there is only mutually assured destruction the only winning move is not to play. Wargames Obviously this lesson cannot be ascribed to our modern world, where we are completely dependent on massive interconnected systems designed to squeeze all the joy and freedom out of life and turn them into profitable numbers while having us make ‘the right’ decisions (i.e. buying the things advertised to us). Tech Start-Ups Are Fine Though, Right? And that brings us to the modern tech start-up, especially those that preach decentralisation and autonomy. Some, I’m sure, are sincere - it’s a statistical impossibility that all of them are presenting a smokescreen, but it’s definitely a minority. There’s a whole rainbow of -washings (green, pink, etc.) that companies use to publicly show off their credentials, whether social, environmental, diversity, or otherwise, while carrying on as normal behind the scenes, but I’m unsure what colour would apply to autonomy. VC-backed start-ups are funded by investment. Investors are looking for a return on their investment - usually by selling the company to bigger investors later on. Even if the first investors strongly believe in the ideals of the company rather than just the opportunity for profit, success and a sale dilutes everything further and further from those initial aims. Driving and supporting autonomy and decentralisation of power inevitably builds systems where it is harder for a single entity to extract value in the form of profit for themselves. Decentralised systems are resilient, not robust. They’re sustainable, not efficient. They tend towards value-sharing, not profit-making. This is a problem if the key incentives that drive an organisation are monetary returns on investments. What’s the Answer? Is there anything that can be done? Well, to a degree. The system is clearly built in such a way that gathering the resources to change it requires a level of investment which makes it hard to then change it without causing yourself damage. Taking part in these larger systems constrains our choices and autonomy, while opting out of them leaves us with ideological movements. Movements which are often unable to take effective actions, and easy for coordinated efforts to subvert for their own ends. At some point there may be enough smaller entities which aren’t profit-driven pulling in the same direction to make a difference, but until then all we can do is just not play the game. Instead of pursuing the big pay cheques of major tech companies, look to smaller organisations with different motivations. A question I was always told to ask when going to interviews was “why are you in business?” If the answer was “to make money”, it was not a sustainable place to work. For an organisation money should be a means to an end. Treating it as an end in itself is just chasing after a high score that means nothing outside the arcade (80s reference just to tie back to there). Wargames Take a lesson from the Luddites. Yes, they failed. Yes, they were utterly crushed by a system that turned people into nothing more than components in a money-making machine with no autonomy, no power, and no identity beyond ‘worker’. Failure doesn’t mean they were wrong. This article should not be taken as an incentive to or an endorsement of taking direct destructive action against datacentres, infrastructure, technology companies or similar. The author fully endorses the ideals of the Luddites, not their methods, or at least not the methods that will lead to similar reactions to those they faced.

Luddites, WarGames, and the Lost Battle for Autonomy

) continues to develop as a transformative force across many spheres of life, already beginning to revolutionize industries and reshape the way we live and work. The topic of AI in warfare will require increasing attention from governments, policymakers, and international organizations. A large part of this is due to significant advancements in the development of autonomous weapons systems (AWS), which use algorithms to operate independently and without human supervision on the battlefield. More broadly, AI in its many forms has the potential to enhance a range of military activities, from the likes of robotics and weaponry to intelligence gathering and decision-making. Artificial Intelligence (AI With such a diversity of potential applications comes a unique set of ethical dilemmas. The benefits of AI in warfare are increased precision, reduced human casualties, and even deterrence against entering into armed conflict in the first place akin to the threat of nuclear war. However, this would mean giving machines the ability to make deliberate life-and-death decisions, blurring the lines of accountability, and possibly going against the fundamental principles of morality in warfare. A Brief Overview of AI in Warfare As the Stockholm International Peace Research Institute outlines, AI has become a crucial part of military strategies and budgets, contributing to the wider ‘arms race’[1]. Combined with the likes of nuclear and atomic threats, geopolitics must therefore question the ethics of the continued weaponization of technology. Some believe that these advancements will ultimately lead to zero-sum thinking dominating world politics. This logic is not new; Alfred Nobel hoped the destructive power of dynamite would put an end to all wars[2]. AI has already started to be incorporated into warfare technology such as in drone swarms, guided missiles, and logistical analysis. Autonomous systems have been incorporated into defensive weaponry for even longer, e.g. antivehicle and antipersonnel mines. Future developments will continue to aspire to increasing levels of autonomy. The US is testing AI bots that can self-fly a modified version of the F-16 fighter jet; Russia is testing autonomous tanks; and China too is developing its own AI-powered weapons[3]. The goal is to protect human life by continuing to mechanize and automate battlefields. “I can easily imagine a future in which drones outnumber people in the armed forces pretty considerably[3]” said Douglas Shaw, senior advisor at the Nuclear Threat Initiative. So, instead of soldiers being deployed on the ground we saved lives by putting them in planes and arming them with missiles. Now with AI, militaries hope to spare even more human life from its forces. Moral Implications of AI in Warfare This sounds great so far. Save lives by using AI to direct drones. Save lives by using AI to launch missiles. The difference between this technological jump in warfare and past innovations is the lack of human input in decision-making. With AWS and lethal autonomous weapons systems (LAWS), we are handing the power to kill a human being over to an algorithm that has no intuitive humanity. Several ethical, moral, and legal issues arise here. Is it fair that human life should be taken in war without another human being on the other side of that action? Does the programmer of an algorithm in a LAWS have the same responsibility in representing their country as a fighter pilot, and/or the same right to contribute to taking enemy life? Like with the ethical dilemmas surrounding autonomous vehicles[4], is it morally justifiable to delegate life-and-death decisions to AI-powered algorithms? From a technological point of view, this will come to depend in part on the transparency of the programming of AWS: training, datasets used, coded preferences, and errors like bias, in these models. Even if we reach an adequate level of accuracy and transparency, should AWS and LAWS be considered moral in warfare? Moral Implications of Just War Theory Just War Theory, credited to St Augustine and Thomas Aquinas in the 13th century[5], evaluates the morality of warfare and ethical decision-making in armed conflict. Across guidelines for (justice of war) and (justice in war), the most notable considerations are: jus ad bellum jus in bello Proportionality: The use of force must be proportional to the objective being pursued and must not cause excessive harm or suffering relative to the anticipated benefits. Discrimination: Also known as non-combatant immunity, this principle requires that combatants distinguish between combatants and non-combatants, and only target the former while minimizing harm to the latter. It could be argued that the use of AI-powered weapons and LAWS do not guarantee adherence to these conventions. On proportionality, AI-backed weaponry would possess the ability to deliver force with greater speed, power, and precision than ever before. Would this level of force necessarily match the threat posed/military objective, especially if used against a country with less technologically advanced weaponry? Similarly, what if a LAWS is fed erroneous intel, or hallucinates and creates an inaccurate prediction? This could lead to the formation and execution of unnecessary military force and disproportionate actions. On the point of discrimination, these technologies are not 100% accurate. When firing a missile at an enemy force, what happens if facial recognition[6] technologies cannot distinguish civilians from combatants? This would undermine the moral distinction between legitimate military targets and innocent bystanders. Case Study A Panel of UN Experts reported the possible use of a LAWS - STM Kargu-2 - in Libya in 2020, deployed by the Turkish military against the Haftar Affiliated Forces (HAF)[7]. Described as being “programmed to attack targets without requiring data connectivity between the operator and the munition” [8], the drone units were eventually neutralized by electronic jamming. The involvement of this remote air technology though changed the tide for what had previously been “a low-intensity, low-technology conflict in which casualty avoidance and force protection were a priority for both parties”[7]. While causing significant casualties, it is not clear whether the unmanned attack drones caused any fatalities[8]. Still, it highlights issues with unregulated, unmanned use of combat aerial vehicles and drones. HAF units were not trained to defend against this form of attack, had no protection from the aerial attacks (which occurred despite the drones being offline), and even in retreat continued to be harassed by the LAWS. This alone begins to breach the principle of proportionality, and even more so when considering that the STM Kargu-2s changed the dynamic of the conflict. Reports go so far as to suggest that “the introduction by Turkey of advanced military technology into the conflict was a decisive element in the… uneven war of attrition that resulted in the defeat of HAF in western Libya during 2020”[7]. International Cooperation and Regulation of AI in Military Applications Since 2018, the UN Secretary-General António Guterres has maintained that LAWS are both politically and morally unacceptable[9]. In his 2023 New Agenda for Peace, Guterres has called for this to be formalized and actioned by 2026. Under this, he suggests a complete ban on the use of AWS which functions without human oversight and does not comply with international law, and regulation of all other AWS. This type of international cooperation and regulation will be necessary to help overcome the ethical concerns we have discussed. For now, the use of AWS without human oversight will cause the most immediate issues. The lack of a human decision-maker creates issues of responsibility. Without a chain of command who takes responsibility for the malfunctioning or general fallibility of an AI-powered system? Moreover, there would be an ensuing lack of accountability. Especially in traditional warfare where there are defined moral principles like the Just War Theory, here there would be no culpable agent for actions taken by autonomous systems. Finally, while there are benefits to increasingly adopting AI in military applications, how these technologies end up being used will define whether it becomes a utopic solution or a proliferation of the already politically destabilizing arms race. Therefore, continued discussion around international, legally binding frameworks for ensuring accountability in AI warfare will arguably be one of the most crucial areas of AI regulation in the near future. References Weaponizing Innovation: Mapping Artificial Intelligence-enabled Security and Defence in the EU War and Technology - Foreign Policy Research Institute How AI Will Revolutionize Warfare Ethical AI and Autonomous Vehicles: Championing Moral Principles in the Era of Self-Driving Cars | HackerNoon Just War Theory | Internet Encyclopedia of Philosophy Bias in Facial Recognition Tech: Explore How Facial Recognition Systems Can Perpetuate Biases | HackerNoon Libya, The Use of Lethal Autonomous Weapon Systems | How does law protect in war? - Online casebook The Kargu-2 Autonomous Attack Drone: Legal & Ethical Dimensions - Lieber Institute West Point Lethal Autonomous Weapon Systems (LAWS) – UNODA

The Ethics of AI in Modern Warfare: Balancing Innovation with Moral Accountability

আধুনিক যুদ্ধে এআই-এর নীতিশাস্ত্র: নৈতিক জবাবদিহিতার সাথে উদ্ভাবনের ভারসাম্য

Die Ethik der KI in der modernen Kriegsführung: Innovation und moralische Verantwortung in Einklang bringen

La ética de la IA en la guerra moderna: equilibrar la innovación con la responsabilidad moral

L'éthique de l'IA dans la guerre moderne : équilibrer l'innovation et la responsabilité morale

आधुनिक युद्ध में एआई की नैतिकता: नैतिक जवाबदेही के साथ नवाचार का संतुलन

現代の戦争における AI の倫理: イノベーションと道徳的責任のバランス

현대전에서의 AI 윤리: 혁신과 도덕적 책임의 균형

A Ética da IA na Guerra Moderna: Equilibrando Inovação com Responsabilidade Moral

Этика ИИ в современной войне: баланс между инновациями и моральной ответственностью

Modern Savaşta Yapay Zeka Etiği: İnovasyonu Ahlaki Sorumlulukla Dengelemek

Đạo đức của AI trong chiến tranh hiện đại: Cân bằng giữa đổi mới với trách nhiệm đạo đức

现代战争中的人工智能伦理：平衡创新与道德责任

When building a generative AI application that needs to call a large language model (LLM) multiple times to complete a task, a common problem is that repeated queries to the LLM can be both expensive and unpredictable. Large models such as GPT-3.5/4 are incredibly resource-intensive to train and run inference; this is reflected in the API charges as well as occasional disruptions to service. ChatGPT was originally released as a research preview and wasn’t intended to be used for production applications. However, its usefulness across a vast array of applications is indisputable, so has exploded. interest in LLMs Since the inception of ChatGPT, users have been looking for ways around the lack of privacy and the inability to control uptime or inference settings when using GPT. This has led to the popularity of free, public models such as Meta’s and later the creation of quantized and lower-parameter versions of Llama that can run on consumer hardware. These public models are capable of providing much of the same functionality as GPT for much less computing power, though at the cost of fewer parameters and fewer verbose outputs. Llama 2 If your application doesn’t necessarily depend on processing excessively large contexts or producing verbose outputs, then hosting your own inference on instances you control can be a more cost-effective option. And when it comes to real-world applications of (RAG), the cost differences may become even more significant. retrieval augmented generation I’ll demonstrate a simple method for combining vector stores, lexical search, and prompt engineering to perform accurate RAG on commodity hardware. Using this method, you can both reduce the complexity of large volumes of information and make running generative AI applications more accurate, efficient, and cost-effective at scale. By using RAG on specific stores of information, you can gain the ability to eliminate hallucinations and create effective and knowledgeable agents from any source material without needing to pay for third-party APIs. To get started, you will need either a DataStax Enterprise 7 instance or to store the vectors and text data, as well as an LLM and a sentence transformer model to generate responses and encode your data with vectors. Depending on the complexity of your data or the user prompts, you may also consider combining this with a DataStax Enterprise 6.8 database that can perform searches to match wider ranges of data, which is what I used in this example. DataStax is continually working on improvements to enable all of these operations with a single database, but for now, I use two databases. DataStax Astra DB Solr Solving for Hallucinations Regardless of which LLM you choose, they all still suffer from . For now, that limitation needs to be resolved by feeding truthful information into the context of prompts to the LLM, otherwise known as RAG. The method by which you locate your information and transform it for the prompts completely depends on your data model, but you can find more pertinent information in a more efficient way by using vector databases. hallucinations Say, for instance, that you have a collection of e-books on a subject you’d like to explore, such as how to play . Under normal circumstances, it would take years to read through the supporting literature and gain enough gameplay experience to reach an expert level. Warhammer 40,000 A targeted question such as, “What can you tell me about Morvenn Vahl of Adepta Sororitas?” would be best answered by a veteran player or any employee at a Warhammer store. While ChatGPT can answer many questions about the game, it unfortunately has no training data that covers this particular character: Compare this to a Llama 2 13B parameter LLM hosted on a consumer workstation with an Nvidia RTX A4000 graphics card. Similarly, the model can demonstrate basic knowledge of the Warhammer universe, but because of the tuning, the model doesn’t care that the character isn’t found and provides instead a best-effort hallucination: If you want to build a chatbot that can help both newcomers and veterans play Warhammer 40,000, then these outputs are unacceptable. To be an effective game guide, the chatbot needs to know the rules of the game, the rules for each unit, some bits of the lore, and some strategy and commentary. Luckily, all of that information on the 10th-edition rules is available for free from Games Workshop and fan websites, and all you need to do is make it searchable to your chatbot app. Compare this to the same 13B Llama model, where with RAG, it is asked to compare a couple of sources on Morvenn Vahl and devise a relevant answer based on the user prompt. This time, the chatbot has access to a database full of all the public information on how to play Warhammer 40,000, 10th Edition: search database and a vector What a difference! Not only does it find pertinent information on this niche character, but it also keeps its outputs in line with the context of how to play the game with the 10th-edition rules. The hardest part in all of this is performing an effective search to find the relevant pages to feed into the LLM. This is where vector databases can be particularly useful. Applying vectors In this example, we’ll use DSE 7 and DSE 6.8 running in Docker instances to satisfy the database requirements of the chatbot application, which needs to be able to compare vectors and perform lexical searches. DSE 7 and Astra DB have introduced the ability to store vectors and perform vector searches, as well as filtering by text matches. We only need to search a few dozen books for this example, so running DSE instances in Docker will be sufficient for most consumer hardware. Using vectors in your databases will help to find documents that are similar to a given query, or they can be used to compare results retrieved from another search. This can help you to overcome the limitations of lexical search and to improve the effectiveness of data models. For instance, something like e-book PDFs can benefit from being encoded with sentence transformers like , and the vectors can be used to run a similarity comparison between a query and a given source. In this case, a sentence transformer model is used to create embeddings of a page’s text in an e-book, and this can enable you to compare to the user’s prompt to figure out if a result is relevant to the query. Relevant pages should contain one or more instances of terms that are similar to the user’s query and result in better similarity scores from the model’s standpoint. miniLM That said, the vectors are best applied as a supplement to an existing lexical search model. If you search by vectors only, then you might end up unexpectedly retrieving unrelated documents and providing them as context where they don’t apply. In this example, the query “What can you tell me about Morvenn Vahl of Adepta Sororitas?” can be transformed by an LLM to a set of simple search terms: Morvenn, Vahl, Adepta, Sororitas The first step in finding relevant documents would be to search for documents that contain those basic terms. This can be done by first filtering for text matches in the database to find keywords in the page text matching such a query. The reason for using an LLM to generate keywords is to provide a wider range of possible keywords to search, as it often attempts to add more keywords that are related but are not in the text of the original prompt. Be careful with this, however, as LLMs can also generate special characters and odd sequences that you will need to sanitize. Once you have at least one result, you can vectorize the user’s query and compare it to the vectors of the lexical search, creating scores of how relevant each result is. This allows you to check the search results for accuracy to the query and set a threshold for rejecting unrelated results when it comes to finally presenting your results to the LLM. In this case, the first step should match pages that specifically show Morvenn Vahl’s index card or gameplay mechanics because those describe the character’s unit in terms of how it plays in the game. If the page meets a certain relevance threshold to the user query, determined by the application, then it gets summarized and placed in a list of results. Finally, the search results can be compiled into a list and fed back to the LLM, where it is asked to use the most relevant contexts to answer the original query. Here is a visualization of the flow: As you can see, the LLM gets called quite frequently for this flow. The LLM is responsible for transforming the user prompt into keywords, summarizing applicable results, and choosing which context best answers a query. Each source to check adds another LLM invocation, which can be quite expensive when making queries to GPT. But if you already have the information you need and just want to summarize it or transform it, then you may not need to use such a large model. In fact, switching to smaller models can provide a number of benefits. By using a smaller LLM, you can reduce the computational cost of each query, which can lead to significant savings over time. This can also result in faster response times for your users, which can improve their overall experience. In this example, where RAG is performed using a small LLM and small databases, all hosted on the same GPU instance, it takes about 30 seconds to retrieve 15 sources, analyze them for relevance, and provide a final answer. And the shorter the prompts (sources), the faster the outputs can be returned. Additionally, this method allows for increased security and scalability. With prompt engineering and a pipeline of calls to the LLM, you’re in is accessed and what the users will get in their responses. In terms of resource usage, the example 13B parameter model only consumes a little over 8GB of VRAM and still provides relevant answers. Depending on needs, this shows potential for even running RAG on myriad other platforms, such as user workstations and mobile devices. full control of how the data Controlling the output Prompt engineering is key to making RAG do exactly what you want. You are in control of how the chatbot interprets the data and the context under which it should be thinking. In this example, we want to ensure that the chatbot knows we are specifically seeking Warhammer information, so we can first ask it to help provide supporting context to the user’s query: Query: “<user query>” Give me a minimal, comma-separated list of Warhammer 40K keywords for a search engine. Respond with only the query. Do not use emojis or special characters. Answer: Warhammer 40,000 is full of terms and names that might appear in other, unrelated popular culture, so it is important to set the context of the RAG in the very first query. This context should be something available to select or modify if your application covers multiple contexts, such as if you need to cover multiple editions of the Warhammer game rules or combine them with the official lore books. Note that the user’s query is always encapsulated with quotes for this experiment. This helps the LLM to distinguish between the query that it’s attempting to directly answer and the separate prompt-engineering instructions, which it must not directly answer. The question/answer portion of the prompt can be adjusted to fit a particular context, but mostly, all you need to be able to do is inform the LLM what it should and should not directly respond to and how to respond. In this case, it’s safe to assume that the LLM has a general knowledge of the game universe since the series is reasonably popular and general information is available for free. The output of this first query helps generate some keywords to use in lexical search without us having to build a cheat sheet into our application. The lexical and vector comparisons can then be performed in the background, and a list of results is compiled for review by the LLM. Because the user’s original prompt is never directly answered with inference at the first step, the LLM only transforms what is found in a search and can easily be stopped from answering queries outside of its guard rails or knowledge base. If there are relevant results from the search: Query: “<user query>” Review these search results and use them to answer the query. Result 1 Result 2 etc. Answer: If there are no relevant results from the search: Query: “<user query>” Politely tell me you searched but could not find an answer for the query. Answer to the best of your knowledge instead. Answer: For added security, you can completely reject or redirect the request when it cannot be served. Query: “<user query>” Politely tell me you searched but could not find an answer to the query. Instruct me to reach out to the Customer Support team for assistance instead. Answer: You can even make the outputs lengthier by asking for more details. As long as you can fit your source material within the context window, the LLM can transform it for you. Query: “<user query>” Review these search results and use them to answer the query. Be as detailed as possible and cite the sources. Result 1 Result 2 etc. Answer: Limitations The LLM has a limited context window and will fail to process exceptionally large pages of text. Consider placing limits on row size so your and easier for the LLM to process. For example, cutting pages into chunks of about 1,000 characters seems to work well, and try to avoid feeding more than four or five detailed answers into the prompt. data is more manageable The LLM has no memory of a conversation aside from what you can fit in the context window. It is possible to build a permanent store of conversation data, but it is not possible for an LLM to fit excessively large conversations or detailed context into a prompt; there is an upper limit to what it can transform. This means that no matter what, at a certain point, you will notice that the LLM seems to “forget” certain details even when they are provided as context; this is just an inherent limitation of the tool. It is best to rely on it for short conversations only and focus on transforming small amounts of text at a time to minimize hallucinations. Randomness in the LLM can be a problem. Testing and tuning will be necessary to determine which prompts work best for your data set and to find which model works best for your use case. In my testing with a 13B parameter model, there was a lot of unpredictability regarding which search keywords were generated from the first prompt, especially as the prompt length increased. For best results, stick to shorter prompts. Conclusion In summary, leveraging RAG by combining models allows for more effective finding and sorting of relevant results and generating agent outputs that are far less prone to hallucinations. The smaller the searchable context, the more precise and accurate the responses. Constructing your own custom pipeline of LLM calls offers far more flexibility in tuning responses toward your desired level of accuracy and guard rails. vector and lexical search While it cannot process excessively large amounts of data within the limited context window, it does offer the ability to create effective assistants on limited knowledge bases, as well as run more concurrent agents on the same or lesser hardware than before. This could open up more possibilities for virtual assistants for applications such as tabletop gaming or even cover more complex topics for use by government, legal and accounting firms, scientific research, energy, and more. If you’re ready to get started building, you can . Create your database and start loading your RAG sources today, with no cloud or database Ops experience required. try Astra DB for free By Mario Charnell-Delgado, DataStax Also published . here

Want to Learn Warhammer 40K? Build This Chatbot with Vectors and RAG on Commodity Hardware

Warhammer 40K শিখতে চান? কমোডিটি হার্ডওয়্যারে ভেক্টর এবং RAG দিয়ে এই চ্যাটবট তৈরি করুন

Möchten Sie Warhammer 40K lernen? Erstellen Sie diesen Chatbot mit Vektoren und RAG auf Commodity-Hardware

¿Quieres aprender Warhammer 40K? Cree este chatbot con vectores y RAG en hardware básico

Vous voulez apprendre Warhammer 40K ? Construisez ce chatbot avec des vecteurs et RAG sur du matériel de base

क्या आप वॉरहैमर 40K सीखना चाहते हैं? कमोडिटी हार्डवेयर पर वेक्टर और आरएजी के साथ इस चैटबॉट का निर्माण करें

ウォーハンマー 40K を学びたいですか? Vectors と RAG を使用してコモディティ ハードウェア上でこのチャットボットを構築する

워해머 40K를 배우고 싶으십니까? 상용 하드웨어에서 벡터와 RAG를 사용하여 챗봇 구축

Quer aprender Warhammer 40K? Construa este chatbot com vetores e RAG em hardware de commodities

Хотите изучить Warhammer 40K? Создайте этого чат-бота с помощью Vector и RAG на стандартном оборудовании

Warhammer 40K'yı Öğrenmek İster misiniz? Ticari Donanımda Vektörler ve RAG ile Bu Chatbot'u Oluşturun

Bạn muốn học Warhammer 40K? Xây dựng Chatbot này với Vector và RAG trên phần cứng hàng hóa

想学习战锤40K吗？在商品硬件上使用向量和 RAG 构建此聊天机器人

**CAYMAN ISLANDS, Cayman Islands, January 30th, 2024/Chainwire/--**The QRDO Foundation, dedicated to the growth and development of the QRDO ecosystem, has announced a strategic partnership with EQ LAB, a leading blockchain developer lab, to launch the Warden Protocol. This initiative accelerates the upcoming launch of the Warden Protocol, an intent-based interoperability protocol built on Cosmos and based on the Fusionchain primitive. Introducing The Warden Protocol The is a modular intent-centric blockchain built on the Cosmos-SDK. At its most basic level, it enables users to create Spaces and Wallets on various blockchains and govern their activity through on-chain intents. Warden Protocol Complex multi-leg transactions can be built, and cross-chain activity protected via complex intents, all enforced on-chain by the Warden Protocol. For the builders in the space, Warden enables smart contracts to be deployed on Cosmos using Solidity and WebAssembly, and facilitates a modular marketplace of key management solutions, from HSM solutions to multi-party computation providers. “The Warden Protocol is the next evolution in intent and interoperability primitives,” said a representative for the QRDO Foundation, “Joining forces with the EQ LAB team enables us to achieve this ambitious vision and ensures that QRDO token holders see the realization of a truly decentralized and open, intent-centric interoperability and key management protocol.” Various applications have pledged their support to deploy on Warden, including: , a liquid staking hub for Cosmos that puts existing liquid staking tokens to work; EQ.finance WARDD, a USD-pegged decentralized stablecoin that provides Warden users with instant access to dollar liquidity; , a pluggable protocol for decentralized funding pools that enable margin trading on any spot DEX; and Marginly SpaceWard, a SAFE-like platform for wallet management and governance. EQ LAB will form the core contributor team to the Warden Protocol and will bring an extensive team of 15 core developers to the Warden Protocol. “We are very pleased to be helping to establish the Warden Protocol as Core Contributors”, said Alex Melikhov, founder of EQ LAB, “as an experienced team of qualified blockchain developers, we see a bright future in the Cosmos ecosystem, and we look forward to seeing incredible value unlocked for both the existing QRDO and Q token holder communities”. WARD Token The Warden Protocol plans to introduce the WARD token via a fair launch mechanism. This launch will take place without any pre-mine or investor allocations, with eligibility for both an airdrop and an incentivised WARD-swap extending first to existing QRDO holders. As a nod to the public goods done by various other Cosmos chains, TIA and ATOM stakers will also be eligible, as well as builders and users from other complimentary protocols and chains. Further details will be announced shortly. The Alfama testnet will go live in the coming weeks, with details for an incentivised testnet to be published shortly. Contact Information Warden Protocol Warden Protocol is a new modular intent-based blockchain based on CosmosSDK that allows users and builders to explore and build cross-chain securely. pr@wardenprotocol.org www.wardenprotocol.org About EQ LAB EQ LAB is an experienced software engineering house specializing in permissionless applications and protocols, acting as a technology partner in a plethora of projects on Cosmos, Ethereum, Arbitrum, Polkadot, and other blockchain platforms. hello@eqlab.io www.eqlab.io About QRDO Foundation: QRDO Foundation has been established to focus on accelerating the adoption of open custody, security and interoperability protocols and nurture the value of the QRDO ecosystem. Contact PR Team QRDO Foundation pr@qrdo.org This story was distributed as a release by Chainwire under HackerNoon’s Business Blogging Program. Learn more about the program here.

QRDO Foundation and EQ LAB Announce Strategic Partnership To Launch The Warden Protocol

As a runner up in HackerNoon's prestigious Startups of the Year 2023 competition, and its CEO have been garnering significant attention in the tech world. Here’s the interview with Mr. Miro to discuss Wallet Factory's meteoric rise as a leading provider of digital wallet solutions, the key factors behind the company's impressive growth, and his vision for the future of the payments industry. Fresh off the heels of this major achievement, the CEO opened up about the company's foundational principles, his personal leadership style that has driven their product innovation and market success, and why he believes Wallet Factory has cemented itself as a startup to watch in the ever-evolving fintech ecosystem. Wallet Factory Mikhail Miro https://hackernoon.com/startups/europe/europe-warsaw-poland Tell us how your startup is changing the world. Wallet Factory is a one-stop embedded finance solutions provider for ambitious, forward-thinking businesses across developing markets in Africa, MENA, and LatAm. The company helps capitalize on innovative FinTech opportunities like digital wallets, personalized loyalty rewards, B2B payment services, financial marketplace, and open banking. Our mission is to facilitate the transition from a "Cash Heavy" to a "Cash Lite" society, aiming to include unbanked citizens into the financial system through our clients. We offer a unique 3-in-1 digital wallet platform that enables banks to serve the unbanked and underbanked segments, fostering financial inclusion and accessibility for those who have not yet had access to traditional banking services. What sets you apart from the competition? We have succeeded in developing all-purpose digital wallets that go bundled with personalized loyalty rewards. Our innovative Wallet and Rewards platforms not only provide digital financial services but also instill long-lasting customer loyalty to engage and retain earlier and newly acquired audiences. Through personalized rewards and secure and user-friendly interfaces, we strive to make payment services easily accessible and attractive, even for non-tech-savvy and picky end customers. What do you love about your team, and why are you the ones to solve this problem? They are the most fantastic people I’ve ever worked with! Wallet Factory's team is comprised of dedicated professionals with a shared passion for enabling financial inclusion through digital financial services. The team's collective expertise and experience in the financial industry, coupled with their commitment to creating innovative FinTech solutions, uniquely position Wallet Factory to address the challenge of fostering financial accessibility and inclusion for unbanked and underbanked populations. If you weren’t building your startup, what would you be doing? Personally, I would likely be working in a leading position in the international banking field, continuing to contribute to the financial industry and helping people access digital financial services. Speaking of Wallet Factory, if it wasn't focused on its current mission, the company would likely be engaged in developing and providing other innovative FinTech solutions to address various challenges in the financial industry, with a continued emphasis on fostering financial inclusion and accessibility for underserved populations. At the moment, how do you measure success? What are your core metrics? As of today, our executives and analysts measure success by the number of end customers using our digital wallet and loyalty rewards platforms, as well as the growth of our clients' revenue streams. Our core metrics include user engagement, customer retention, and the expansion of our services in emerging markets. What goals are you looking forward to accomplishing in 2024? In 2024, we aim to further expand our presence in emerging markets and help more banks and non-banking financial institutions serve the unbanked and underbanked segments. We also plan to enhance our digital wallet and loyalty rewards platforms with new features and integrations to better meet the needs of our clients and their customers. Which trend(s) are you most excited about in 2024? Share your reasons. We are excited about the continued growth and adoption of digital wallets and mobile payment services, as well as the potential for increased collaboration between banks and FinTech companies to drive financial inclusion and accessibility. 2023 had been another crazy year, especially in tech, with layoffs and the Generative AI takeover! Which trend are you most concerned about? Be as brief or as detailed as you like. We are most concerned about the potential for increased competition in the digital wallet and FinTech space, which could lead to a saturated market and challenges in differentiating our offerings from those of other companies. One of the most promising trends I’d like to mention is the adoption of AI, as you have deftly put it in your question. I’m definite about the fact that this trend will likely bring more personalized solutions to end customers in 2024, driving greater user experiences and brand loyalty. Share your biggest success so far and/or your biggest failure so far. Our biggest success so far has been the launch of over 25 revenue-generating eWallet and customer loyalty projects in 12 countries across 5 continents, with over 4.5M end customers worldwide. One of the biggest challenges our team has ever experienced was the difficulty in penetrating certain markets due to ongoing regulatory challenges and the need for more localized solutions. Fortunately, we’ve learned a lot from that lesson and now can easily tailor our digital wallet and loyalty rewards platforms to the business specifics of any industry or region. We would love your feedback on HackerNoon as a tech publication! How has your experience been with us? Our experience with HackerNoon has been more than positive, as we have been able to share our insights and expertise on digital wallets and FinTech trends with a wide audience of tech enthusiasts and businesses. Sure thing, Wallet Factory will be looking forward to taking part in any further tech competitions of this kind! Any words of wisdom you’d like to share with us? My piece of advice for businesses and tech enthusiasts is to stay informed in today’s ever-changing world while keeping tabs on the latest trends and developments in the digital wallet and FinTech space. Much like we care about our clients’ success, businesses should think first about bettering their customers’ experiences with the brand. It can help you be prepared to adapt and innovate efficient solutions across your business operations in order to stay ahead of the competition and satisfy your customers. Startups of The Year is HackerNoon’s Flagship community-driven event celebrating startups that survived and thrived in 2023. 30,000 startups across 4200+ cities and six continents participated this year to be crowned the best startup in their city. See our global winners announcement . here Companies that blog more get 67% more leads - Jumpstart your company’s blogging journey by purchasing our . Business Blogging Package The 2023 Startups Of The Year is sponsored by: . . The winners will get a free .Tech Domain, a HackerNoon NFT, and a Tech Company News Page. TECH Domains

Meet Wallet Factory, Runner-up of the Startups of the Year in Warsaw, Poland

Many people have been worried about the safety concerns that Artificial Intelligence and modern technologies pose to human society. We have thousands of movies about how AI and humanoid robots take over the world and make humans their slaves. There is also a popular thought experiment dedicated to it. It is known as Roko's basilisk. "Roko's Basilisk is a thought experiment that states that an otherwise benevolent in the future would be incentivized to create a virtual reality simulation to torture anyone who knew of its potential existence but did not directly contribute to its advancement or development, to incentivize said advancement." artificial superintelligence (AI) These are mere speculation, and at the current period and in the next decade, we will have a technology that will endanger our existence. But, I fear what I just said would be proven wrong if we keep ignoring the necessity of guidelines to regulate AI technologies and their development. ChatGPT maker OpenAI has recently updated its Terms of Service. It has altered its usage policy to get rid of a prohibition on using its technology for "military and warfare." Previously, one of the leading tech companies prohibited its tools and services specifically to use for "weapons development, military and warfare," but this was in action only before . January 10 of this year OpenAI has now made changes to its policy to only disallow the use that would "bring harm to others," IT magazine Computer World first reported. "Our policy does not allow our tools to be used to harm people, develop weapons, for communications surveillance, or to injure others or destroy property," a spokesperson from OpenAI reportedly said. "There are, however, national security use cases that align with our mission. For example, we are already working with DARPA to spur the creation of new cybersecurity tools to secure open-source software that critical infrastructure and industry depend on. It was not clear whether these beneficial use cases would have been allowed under ‘military’ in our previous policies. So the goal with our policy update is to provide clarity and the ability to have these discussions," he added. This sudden change will now allow the company to work closely with the US military and possibly with the militaries of US allies. This will write a new chapter on the development of AI technologies and their potential uses in modern warfare and other military tools. There are also reports that certain people and employees are skeptical of this new development and the future of working closely with the military. Let's keep the negativities aside and discuss the positive aspects of using AI and robots to fight wars for us. Fights will be fair Future wars would be fought by robots. A nation can demonstrate its strength and technological superiority. There would be less chance of foul playing when an international convention is signed for the ethics of robot fighters in wars. There will be a significant decrease in human lives Hundreds of thousands of brave patriots lose their lives in battles. We need people like them. We need their dedication and their love for the motherland. Humans killing other humans over petty issues like land and resources sounds so ridiculous when you look at it from a general perspective. This is true we can save countless lives by using robots to fight for us. We don't even need to wage out large-scale battles when we can just make the best fighters from both the conflicting sides to fight a death match and decide the outcome in the favor of the winning side. Improved Accuracy and Remote Operations capabilities Artificial fighters can take into account the direction of the wind, air pressure, temperature, muzzle velocity, drag coefficient, and various other factors, which can improve accuracy significantly. They can also be used for remote operations and can be controlled from hundreds of kilometers away with state-of-the-art technologies. They can be replaced quickly, can respond rapidly in times of emergency, and most important of all, the humans will not feel the emotional impact of losing comrades, and the mothers and sisters will not shed any more tears. The sons and daughters will not feel alone anymore. That is enough talk of the future. Some experts had already anticipated the increase in interest of the Federal Government and the US military in AI technologies in response to the Chinese authorities facilitating AI technologies developments and incorporating them into the Chinese military. It is sad to see that the Government is prioritizing AI developments to compete with rivals when there is a strong need for an international convention to regulate this fascinating yet destructive technology. Also published . here

AI in Warfare: OpenAI's Policy Shift Regarding Military Usage of Its Tools

The has given an exceptional momentum to the cybercrime. released a cybersecurity report in 2022 on a global scale. The report highlighted that the number of cyber attacks worldwide increased by half. Russian invasion of Ukraine CheckPoint Hackers primarily targeted international economic sectors, with over . Governmental, financial, and media sectors faced over attacks per week. 600 attacks per day two and a half thousand Besides, the time and money losses due to cyber breaches are outrageous. In the US, the cost of just a single is a whopping That’s insane! data breach $9.23 million. Today, we are living in a world where keyboards are mightier than swords, and firewalls play the role of digital fortresses. A world where sensitive information is stolen, critical infrastructure is hacked, and wars are fought online. Who would've thought‍? Let’s dive into the world of cyber warfare and see how devastating it can be for nations across the world. What Is a Cyberwar? Reading the word “war,” you might have visualized generals, weapons, trenches, and alike. But it’s a different kind of war. It's a digital dance of deception and disruption, where the battlefield isn't marked by trenches but by lines of code. Here, the generals wear hoodies, and their weapon of choice is a well-worn keyboard. Cyber warfare is a form of internet attack directed at a country's government, nation, or military infrastructure with the aim of causing disruption and inflicting damage. Cyber warfare should not be confused with the terrorist use of cyberspace or with cyber espionage and cybercrime. In comparison to cybercrime, cyber attacks, and cyber terrorism: Cyber warfare pertains to the conflict itself, whereas a cyber attack is associated with the methods and strategies employed. Cyber terrorism involves the "politically motivated use of computers and information technologies to cause serious disruptions or widespread fear in society." Cyber warfare differs from cyber terrorism in that it represents organized efforts by a national government to conduct operations in cyberspace against foreign states. So, don’t assume yourself a soldier the next time your YouTube channel gets hacked because a cyberwar is a digital conflict between two nations, not individuals. This means it excludes many other forms of attacks that may be misunderstood as cyberwar. For instance, if a cyber attack is initiated by an individual hacker or a group, it cannot be called cyber warfare. However, if those hackers are backed by the state, then it comes under a cyberwar. The main objective of every cyber war is to disrupt, damage, or destroy critical systems. So it may take many forms, such as: Attacks on financial infrastructure. Attacks on public infrastructure like electrical systems or dams. Attacks on safety infrastructure, like early warning systems and target signals. Attacks against military organizations or resources. What Does Cyber Warfare Look Like? Like normal warfare, which can range from minor skirmishes to full-fledged battles, cyberwar varies in severity and target. In most cases, computer systems are not the prime target. They are targeted because they manage real-world systems, such as power grids and airports. Stock prices spiral into chaos as hackers mess with the stock exchange data. The next day, trains stop running because traffic signals do not work. You cannot drive anywhere as traffic lights freeze on red. Sure, it might bring a fleeting moment of joy, a logical excuse to dodge the workplace drama. "Sorry, boss, can't make it today. The traffic lights are having a digital tantrum!" But honestly, this can ruin the whole country if it persists. In no time, a country could be in chaos. Without crossing the boundaries, your enemies can disrupt the whole nation. Do you remember when declared a State of Emergency in 2022 after weeks of a major ransomware attack? Costa Rica Layer 7 Cyber Attacks - The Main Challenge Of The Year At the beginning of the Russian-Ukrainian cyber war, attacks at levels L3 and L4 were mostly used to disrupt infrastructure operations. Attacks at this level simply overload networks or applications, disrupting their bandwidth capacity. However, currently, cyber warfare is conducted with more sophistication. Participants resort to Layer 7 attacks. These are intelligent attacks aimed at finding vulnerabilities in the cyberspace infrastructure, blocking or disrupting its operations for an extended period. Real-Life Examples of a Cyber Warfare We have a few real-life incidents of cyberwar as well. Luckily, there are not many, at least for now. Russia-Ukraine Conflict The conflict between has extended into cyberspace. Various cyber incidents have been attributed to state-sponsored hacking groups. Out of 30 cyberwar incidents between both countries, 28 were initiated by Russia. Russia and Ukraine Russian cyber operations include disrupting Ukrainian television companies before missile strikes, intensifying information campaigns, and attacking private services, such as and . Monobank Kyivstar This coordination of cyberattacks with military operations aimed to disrupt communication and control systems, enhancing the impact of subsequent missile strikes. Ukraine, on the other side, has responded robustly to Russian cyberattacks, despite being a primary target. The Ukrainian cyber army conducted over a million DDoS attacks on Russian infrastructure in 2022, showcasing a resilient opposition to cyber threats. Stuxnet Attack on Iranian Nuclear Facilities marked one of the first instances of a cyber weapon causing real-world damage. It involved a sophisticated worm designed to target Iran's nuclear program. Stuxnet sabotaged Iran's uranium enrichment facilities and caused physical damage to centrifuges. Stuxnet The worm managed to infect The result? This incident ruined around 900 centrifuges. This incident is widely believed to be a joint operation by the U.S. and Israel. revealed that a Dutchman, recruited by AIVD, was behind this sabotage. However, Dutch politicians were unaware of this operation. 20,000 devices in 14 Iranian nuclear facilities. Volkskrant NotPetya Ransomware Attack , a notorious ransomware attack that occurred in 2017, initially targeted Ukraine. But quickly, this attack spread globally, mimicking the COVID-19 of the physical world. It affected major organizations and critical infrastructure worldwide. The victims then spent billions of dollars to recover. NotPetya However, the Russian government denied involvement. Besides, one of the victim’s insurers was linking this incident to cyberspace politics. But they failed to get war-related exclusion and paid as . However, the analyst still connects this incident with the political conflict. $1.4 billion cyber attack coverage The Petya virus on June 27, 2017, paralyzed the operations of companies worldwide. What Are the Types of Cyber Warfare? There are different types of cyber war. Some of them aim to compromise data security. Others leverage computer hacking to threaten human security. Let’s discuss some common types of cyber warfare. Espionage This type of cyberwar attack aims to steal the secrets of other countries. In espionage, a spear-phishing attack or a botnet is used. The intent here is to get access to a target’s computer and extract sensitive information. Isn’t it like a script of a spy movie? After pinpointing sensitive information, organizations assess potential threats to this data. This may involve considering external parties seeking to steal the information. Likewise, there may be competitors aiming for an advantage through data theft. This may also involve internal risks such as disgruntled workers or negligent employees. Distributed Denial-of-Service Attack In such a type of attack, the target website is flooded with fake requests. This is to render the site unavailable for legitimate users. This cyber downpour often targets critical websites, disrupting the virtual lives of military personnel, safety officials, citizens, scientists, and the like. Especially those used by military personnel, safety personnel, citizens, scientists, or alike. In , Russia levied a DDoS attack on Ukraine and crippled its election commission. March 2014 Modern economic systems depend on computers to function. So, the attacking country targets the stock market, payment systems, and banks. They freeze such critical websites with a DDoS attack. Afterward, they infect these devices and steal or manipulate data. Malware and Viruses At the heart of cyberwarfare are malicious software and viruses. These digital weapons are designed to infiltrate computer systems and steal sensitive information. For instance, the notorious Stuxnet attack. Similarly, Russia is using new on Ukrainian targets. These have been installed on several hundred machines in Ukraine. Besides, a Remote Manipulator System (RMS) is being distributed via fake “evacuation plan” emails. Wiper malware The Future of Cyber War So now, you understand what cyberwar is. Such digital attacks can disrupt the whole country. Today, we might not have too many cyber war-related incidents. However, modern technology is pushing the boundaries and making the impossible happen. The Russian invasion in February 2022 has given new momentum to cybercrime. Today, every rich and skilled country is investing in information warfare and cybersecurity. , over 30 countries are developing the ability to launch cyberattacks. However, many of these efforts are kept secret. This secrecy has raised concerns about a hidden cyber arms race. According to US intelligence Besides, the current tension between has also felt the waves of cyberattacks. Both sides have intensified their cyber offensive operations in recent weeks. Iranian state-sponsored hackers have said this clearly on Telegram. They said they hacked critical Israeli systems, like electrical grids. Their warfare mechanism exactly replicated the Russian hacking incident. Iran and Israel warned on Tuesday of a phishing attack by an Iran-based hacking squad. On the other side, also claim that they have 70 percent of gas stations across Iran. Israel’s National Cyber Directorate Israeli hackers There are plenty of devastating stats that scream cyberwar is yet another pandemic this world has to deal with. And what about the financial losses pertaining to cyberwarfare? predicts global cybercrime costs to increase annually by 15%, hitting annually by 2025. Cybersecurity Ventures $10.5 trillion Likewise, the theft of IT data and equipment, as well as industrial and digital espionage and sabotage, will cost Germany a staggering , as per the German digital association. 206 billion euros in 2023 In addition, this involving over 1000 companies estimated that the loss will exceed the mark for the third consecutive year. survey 200 billion euro Final Words So this means that cyberwar is not just a myth anymore; it’s a serious global concern. It will get out of control if no strict actions are taken. International relations and cyber security are interrelated. Whenever the relationship between two nations is disrupted, the threat of cyber warfare will be there. The internet doesn't care about borders; it's like a big open field where threats can run wild. So, countries must work together, wearing their digital cowboy hats, to handle the challenges that pop up.

The Face of Modern Conflict: What You Need to Know About Cyber Warfare

আধুনিক সংঘাতের মুখ: সাইবার যুদ্ধ সম্পর্কে আপনার যা জানা দরকার

Das Gesicht moderner Konflikte: Was Sie über Cyber-Kriegsführung wissen müssen

El rostro del conflicto moderno: lo que necesita saber sobre la guerra cibernética

Le visage des conflits modernes : ce que vous devez savoir sur la cyberguerre

आधुनिक संघर्ष का चेहरा: साइबर युद्ध के बारे में आपको क्या जानना चाहिए

現代の紛争の顔: サイバー戦争について知っておくべきこと

현대 분쟁의 얼굴: 사이버 전쟁에 대해 알아야 할 사항

A face do conflito moderno: o que você precisa saber sobre a guerra cibernética

Лицо современного конфликта: что нужно знать о кибервойне

Modern Çatışmanın Yüzü: Siber Savaş Hakkında Bilmeniz Gerekenler

Bộ mặt của xung đột hiện đại: Những điều bạn cần biết về chiến tranh mạng

现代冲突的面貌：您需要了解的有关网络战的知识

Microsoft v. United States (2016) Court Filing, retrieved on July 14, 2016, is part of . You can jump to any part in this filing . This part is 19 of 22. HackerNoon’s Legal PDF Series here CONCLUSION We conclude that Congress did not intend the SCA’s warrant provisions to apply extraterritorially.  The focus of those provisions is protection of a user’s privacy interests.  Accordingly, the SCA does not authorize a U.S. court to issue and enforce an SCA warrant against a United States‐based service provider for the contents of a customer’s electronic communications stored on servers located outside the United States.  The SCA warrant in this case may not lawfully be used to compel Microsoft to produce to the government the contents of a customer’s e‐mail account stored exclusively in Ireland.  Because Microsoft has otherwise complied with the Warrant, it has no remaining lawful obligation to produce materials to the government. We therefore the District Court’s denial of Microsoft’s motion to quash; we its order holding Microsoft in civil contempt of court; and we this cause to the District Court with instructions to quash the warrant insofar as it demands user content stored outside of the United States. REVERSE VACATE REMAND GERARD E. LYNCH, , concurring in the judgment: Circuit Judge I am in general agreement with the Court’s conclusion that, in light of the presumption against extraterritorial application of congressional enactments, the Stored Communications Act (“SCA” or the “Act”) should not, on the record made by the government below, be construed to require Microsoft to turn over records of the content of emails stored on servers in Ireland. I write separately to clarify what, in my view, is at stake and not at stake in this case; to explain why I believe that the government’s arguments are stronger than the Court’s opinion acknowledges; and to emphasize the need for congressional action to revise a badly outdated statute. Continue Reading . Here About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case No. 15–777 retrieved on September 27, 2023, from is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. cases.justia.com

Microsoft Wins Appeal to Quash Warrant Asking it to Share Customer's Data

Microsoft v. United States (2016) Court Filing, retrieved on July 14, 2016, is part of . You can jump to any part in this filing . This part is 11 of 22. HackerNoon’s Legal PDF Series here 1. Plain Meaning of the SCA As observed above, the SCA permits the government to require service providers to produce the contents of certain priority stored communications “only pursuant to a warrant issued using the procedures described in the Federal Rules of Criminal Procedure (or, in the case of a State court, issued using State warrant procedures) by a court of competent jurisdiction.”  18 U.S.C. § 2703(a), (b)(1)(a).  The provisions in § 2703 that permit a service provider’s disclosure in response to a duly obtained warrant do not mention any extraterritorial application, and the government points to no provision that even implicitly alludes to any such application.  No relevant definition provided by either Title I or Title II of ECPA, see 18 U.S.C. §§ 2510, 2711, suggests that Congress envisioned any extraterritorial use for the statute. When Congress intends a law to apply extraterritorially, it gives an “affirmative indication” of that intent.  Morrison, 561 U.S. at 265.  It did so, for example, in the statutes at issue in Weiss v. National Westminster Bank PLC, 768 F.3d 202, 207 & n.5 (2d Cir. 2014) (concluding that definition of “international terrorism” within 18 U.S.C. § 2331(1) covers extraterritorial conduct because Congress referred to acts that “occur primarily outside the territorial jurisdiction of the United States”) and United States v. Weingarten, 632 F.3d 60, 65 (2d Cir. 2011) (concluding that 18 U.S.C. § 2423(b) applies to extraterritorial conduct because it criminalizes “travel in foreign commerce undertaken with the intent to commit sexual acts with minors” that would violate United States law had the acts occurred in the jurisdiction of the United States).  We see no such indication in the SCA. We emphasize further that under § 2703, any “court of competent jurisdiction”―defined in § 2711(3)(B) to include “a court of general criminal jurisdiction of a State authorized by the law of that State to issue search warrants”―may issue an SCA warrant.  Section 2703(a) refers directly to the use of State warrant procedures as an adequate basis for issuance of an SCA warrant.  18 U.S.C. § 2703(a).  We think it particularly unlikely that, if Congress intended SCA warrants to apply extraterritorially, it would provide for such far‐reaching state court authority without at least “address[ing] the subject of conflicts with foreign laws and procedures.”  Aramco, 499 U.S. at 256; see also American Ins. Ass’n v. Garamendi, 539 U.S. 396, 413 (2003) (describing as beyond dispute the notion that “state power that touches on foreign relations must yield to the National Government’s policy”). The government asserts that “[n]othing in the SCA’s text, structure, purpose, or legislative history indicates that compelled production of records is limited to those stored domestically.”  Gov’t Br. at 26 (formatting altered and emphasis added).  It emphasizes the requirement placed on a service provider to disclose customers’ data, and the absence of any territorial reference restricting that obligation.  We find this argument unpersuasive:  It stands the presumption against extraterritoriality on its head.  It further reads into the Act an extraterritorial awareness and intention that strike us as anachronistic, and for which we see, and the government points to, no textual or documentary support. [20] [20] Seeking additional grounds for its position that to apply Morrison in this case is to proceed on a false premise, the government argues that the presumption against extraterritoriality applies only to “substantive provisions” of United States law, and that the SCA’s warrant provisions are procedural.  Gov’t Br. at 31.  The proposition that the SCA’s protections are merely procedural might reasonably be questioned.  But even assuming that they are procedural, the government gains no traction with this argument, which we rejected in Loginovskaya v. Batratchenko, 764 F.3d 266, 272‐73 (2d Cir. 2014). Continue Reading . Here About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case No. 15–777 retrieved on September 27, 2023, from is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. cases.justia.com

Who in the US Has the Authority to Issue Warrants for Data Stored Overseas?

Microsoft v. United States (2016) Court Filing, retrieved on July 14, 2016, is part of . You can jump to any part in this filing . This part is 12 of 22. HackerNoon’s Legal PDF Series here 2. The SCA’s Use of the Term of Art “Warrant” Congress’s use of the term of art “warrant” also emphasizes the domestic boundaries of the Act in these circumstances. In construing statutes, we interpret a legal term of art in accordance with the term’s traditional legal meaning, unless the statute contains a persuasive indication that Congress intended otherwise.  See F.A.A. v. Cooper, 132 S. Ct. 1441, 1449 (2012) (“[W]hen Congress employs a term of art, ‘it presumably knows and adopts the cluster of ideas that were attached to each borrowed word in the body of learning from which it was taken.’”) (quoting Molzof v. United States, 502 U.S. 301, 307 (1992)).  “Warrant” is such a term of art. The term is endowed with a legal lineage that is centuries old.  The importance of the warrant as an instrument by which the power of government is exercised and constrained is reflected by its prominent appearance in the Fourth Amendment to the United States Constitution: The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized. U.S. Const. amend. IV.  It is often observed that “[t]he chief evil that prompted the framing and adoption of the Fourth Amendment was the indiscriminate searches and seizures conducted by the British under the authority of general warrants.”  United States v. Galpin, 720 F.3d 436, 445 (2d Cir. 2013) (internal quotation marks omitted).   Warrants issued in accordance with the Fourth Amendment thus identify discrete objects and places, and restrict the government’s ability to act beyond the warrant’s purview ― of particular note here, outside of the place identified, which must be described in the document.  Id. at 445–46. As the term is used in the Constitution, a warrant is traditionally moored to privacy concepts applied within the territory of the United States:  “What we know of the history of the drafting of the Fourth Amendment . . . suggests that its purpose was to restrict searches and seizures which might be conducted by the United States in domestic matters.” In re Terrorist Bombings of U.S. Embassies in East Africa, 552 F.3d 157, 169 (2d Cir. 2008) (alteration omitted and ellipses in original) (quoting United States v. Verdugo‐Urquidez, 494 U.S. 259, 266 (1990)).  Indeed, “if U.S. judicial officers were to issue search warrants intended to have extraterritorial effect, such warrants would have dubious legal significance, if any, in a foreign nation.”  Id. at 171. Accordingly, a warrant protects privacy in a distinctly territorial way.[21] The SCA’s legislative history related to its post enactment amendments supports our conclusion that Congress intended to invoke the term “warrant” with all of its traditional, domestic connotations.[22] Since the SCA’s initial passage in 1986, Congress has amended § 2703 to relax some of the Rule 41 requirements as they relate to SCA warrants.  Although some address the reach of SCA warrants, none of the amendments contradicts the term’s traditional domestic limits.  See USA PATRIOT ACT, Pub. L. 107‐56, § 220; 115 Stat. 272, 291–92 (2001) (codified at 18 U.S.C. § 2703(a), (b)); 21st Century Department of Justice Appropriations Authorization Act, Pub. L. 107‐273, § 11010, 116 Stat. 1758, 1822 (2002) (codified at 18 U.S.C. § 2703(g)); Foreign Evidence Request Efficiency Act of 2009, Pub. L. 111‐79, § 2, 123 Stat. 2086, 2086 (2009) (codified at 18 U.S.C. § 2711(3)(A)). These amendments to the SCA are fully consistent with the historical role of warrants as legal instruments that pertain to discrete objects located within the United States, and that are designed to protect U.S. citizens’ privacy interests. The magistrate judge took a different view of the legislative history of certain amendments to the SCA. He took special notice of certain legislative history related to the 2001 amendment to the warrant provisions enacted in the USA PATRIOT ACT. A House committee report explained that “[c]urrently, Federal Rules [sic] of Criminal Procedure 41 requires that the ‘warrant’ be obtained ‘within the district’ where the property is located.  An investigator, for example, located in Boston . . . might have to seek a suspect’s electronic e‐mail from an Internet service provider (ISP) account located in California.” In re Warrant, 15 F. Supp. 3d at 473 (quoting H.R. Rep. 107‐236(I), at 57 (2001)).  The magistrate judge reasoned that this statement equated the location of property with the location of the service provider, and not with the location of any server. Id. at 474. But this excerpt says nothing about the need to cross international boundaries; rather, while noting the “cross‐jurisdictional nature of the Internet,” it discusses only amendments to Rule 41 that allow magistrate judges “within the district” to issue warrants to be executed in other “districts”—not overseas.  Id. at 473 (quoting H.R. Rep. 107‐236(I), at 58).  Furthermore, the Committee discussion reflects no expectation that the material to be searched and seized would be located any place other than where the service provider is located.  Thus, the Committee’s hypothetical focuses on a situation in which an investigator in Boston might seek e‐mail from “an Internet service provider (ISP) account located in California.”  To our reading, the Report presumes that the service provider is located where the account is—within the United States.[23] [21] The government argues that the SCA’s warrant provisions were “modeled after the Right to Financial Privacy Act,” 12 U.S.C. §§ 3402(3), 3406, and that the latter act also “envisions that warrants—along with subpoenas and summonses—will trigger a disclosure requirement.” Gov’t Br. at 19 (citing S. Rep. No. 99‐541, at 3).  It points to no authority definitively construing the latter act’s warrant provisions, however, nor any acknowledgment in the history of the SCA that enforcement of the warrant’s disclosure commands would cross international boundaries. For these reasons, we accord little weight to the observation. [22] We note that a 2009 amendment to Rule 41 expressly authorizes the use of such warrants to seize electronically‐stored data, without abandoning the requirement that the warrant specify the place from which the data is to be seized.  See Fed. R. Crim. P. 41(e)(2)(B) (allowing magistrate judge to “authorize the seizure of electronic storage media or the seizure or copying of electronically stored information” (emphasis added)). [23] Our brief discussion here of the law of warrants is offered in aid only of our interpretation of the statutory language.  Consequently, we do not consider whether the Fourth Amendment might be understood to impose disclosure‐related procedural requirements more stringent than those established by the SCA.  See United States v. Warshak, 631 F.3d 266, 288 (6th Cir. 2010) (finding Fourth Amendment protects certain electronic communications based on users’ reasonable expectations of privacy); see also Email Privacy Act, H. R. 699, 114th Cong. § 3 (passed by House Apr. 27, 2016) (requiring government to obtain warrant before obtaining documents stored online). Continue Reading . Here About HackerNoon Legal PDF Series: We bring you the most important technical and insightful public domain court case filings. This court case No. 15–777 retrieved on September 27, 2023, from is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. cases.justia.com

Stories, News & Mentions about WARDA

Warda’s Stories on HackerNoon

Warda's News & Mentions