Blend-ed opens the door to a revolution in education through an integrated approach, where online and instructor-led modes are effectively blended to achieve the best learning outcomes

Blend-ed

Science and poetry are blended in a happy conjunction

Authors: (1) Savvas Petridis, Google Research, New York, New York, USA; (2) Ben Wedin, Google Research, Cambridge, Massachusetts, USA; (3) James Wexler, Google Research, Cambridge, Massachusetts, USA; (4) Aaron Donsbach, Google Research, Seattle, Washington, USA; (5) Mahima Pushkarna, Google Research, Cambridge, Massachusetts, USA; (6) Nitesh Goyal, Google Research, New York, New York, USA; (7) Carrie J. Cai, Google Research, Mountain View, California, USA; (8) Michael Terry, Google Research, Cambridge, Massachusetts, USA. Table Of Links Abstract & Introduction Related Work Formative Study Constitution Maker Implementation User Study Findings Discussion Conclusion and References 9 CONCLUSION This paper presents ConstitutionMaker, a tool for interactively refining LLM outputs by converting users’ intuitive feedback into principles. ConstitutionMaker’s design is informed by a formative study, where we also collected and classified the types of principles users wanted to write. ConstitutionMaker incorporates three principle-elicitation features: kudos, critique, and rewrite. In a user study with 14 industry professionals, participants felt that ConstitutionMaker helped them (1) write principles that effectively guided the chatbot, (2) convert their feedback into principles more easily, and (3) write principles more efficiently, with (4) less mental demand than the baseline. This was due to ConstitutionMaker supporting their thought processes, including helping them to: identify ways to improve the bot’s responses, convert their intuition into verbal feedback, and phrase their feedback as specific principles. There are many avenues of future work, including supporting users in iterating on and clarifying their principles, organizing larger sets of principles and supporting multiple writers, and helping users test chatbots across multiple user journeys. Together, these findings inform future tools that support interactively customizing LLM outputs. REFERENCES [1] Yuntao Bai, Saurav Kadavath, Sandipan Kundu, Amanda Askell, Jackson Kernion, Andy Jones, Anna Chen, Anna Goldie, Azalia Mirhoseini, Cameron McKinnon, Carol Chen, Catherine Olsson, Christopher Olah, Danny Hernandez, Dawn Drain, Deep Ganguli, Dustin Li, Eli Tran-Johnson, Ethan Perez, Jamie Kerr, Jared Mueller, Jeffrey Ladish, Joshua Landau, Kamal Ndousse, Kamile Lukosuite, Liane Lovitt, Michael Sellitto, Nelson Elhage, Nicholas Schiefer, Noemi Mercado, Nova DasSarma, Robert Lasenby, Robin Larson, Sam Ringer, Scott Johnston, Shauna Kravec, Sheer El Showk, Stanislav Fort, Tamera Lanham, Timothy TelleenLawton, Tom Conerly, Tom Henighan, Tristan Hume, Samuel R. Bowman, Zac Hatfield-Dodds, Ben Mann, Dario Amodei, Nicholas Joseph, Sam McCandlish, Tom Brown, and Jared Kaplan. 2022. Constitutional AI: Harmlessness from AI Feedback. arXiv:2212.08073 [cs.CL] [2] Svetlin Bostandjiev, John O’Donovan, and Tobias Höllerer. 2012. TasteWeights: A Visual Interactive Hybrid Recommender System. In Proceedings of the Sixth ACM Conference on Recommender Systems (Dublin, Ireland) (RecSys ’12). Association for Computing Machinery, New York, NY, USA, 35–42. https://doi.org/10.1145/ 2365952.2365964 [3] Stephen Brade, Bryan Wang, Mauricio Sousa, Sageev Oore, and Tovi Grossman. 2023. Promptify: Text-to-Image Generation through Interactive Prompt Exploration with Large Language Models. In Proceedings of the 35th Annual ACM Symposium on User Interface Software and Technology (UIST). https: //doi.org/10.1145/3586183.3606725 arXiv:2304.09337 [cs.HC] [4] Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel Ziegler, Jeffrey Wu, Clemens Winter, Chris Hesse, Mark Chen, Eric Sigler, Mateusz Litwin, Scott Gray, Benjamin Chess, Jack Clark, Christopher Berner, Sam McCandlish, Alec Radford, Ilya Sutskever, and Dario Amodei. 2020. Language Models are Few-Shot Learners. In Advances in Neural Information Processing Systems, H. Larochelle, M. Ranzato, R. Hadsell, M.F. Balcan, and H. Lin (Eds.), Vol. 33. Curran Associates, Inc., 1877–1901. https://proceedings.neurips.cc/paper_files/paper/2020/file/ 1457c0d6bfcb4967418bfb8ac142f64a-Paper.pdf [5] Sébastien Bubeck, Varun Chandrasekaran, Ronen Eldan, Johannes Gehrke, Eric Horvitz, Ece Kamar, Peter Lee, Yin Tat Lee, Yuanzhi Li, Scott Lundberg, Harsha Nori, Hamid Palangi, Marco Tulio Ribeiro, and Yi Zhang. 2023. Sparks of Artificial General Intelligence: Early experiments with GPT-4. arXiv:2303.12712 [cs.CL] [6] Ernie Chang, Xiaoyu Shen, Hui-Syuan Yeh, and Vera Demberg. 2021. On Training Instance Selection for Few-Shot Neural Text Generation. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 2: Short Papers). Association for Computational Linguistics, Online, 8–13. https: //doi.org/10.18653/v1/2021.acl-short.2 [7] Qiaochu Chen, Xinyu Wang, Xi Ye, Greg Durrett, and Isil Dillig. 2020. MultiModal Synthesis of Regular Expressions. In Proceedings of the 41st ACM SIGPLAN Conference on Programming Language Design and Implementation (London, UK) (PLDI 2020). Association for Computing Machinery, New York, NY, USA, 487–502. https://doi.org/10.1145/3385412.3385988 [8] Katy Ilonka Gero, Vivian Liu, and Lydia Chilton. 2022. Sparks: Inspiration for Science Writing Using Language Models. In Proceedings of the 2022 ACM Designing Interactive Systems Conference (Virtual Event, Australia) (DIS ’22). Association for Computing Machinery, New York, NY, USA, 1002–1019. https: //doi.org/10.1145/3532106.3533533 [9] Sandra G. Hart and Lowell E. Staveland. 1988. Development of NASA-TLX (Task Load Index): Results of Empirical and Theoretical Research. In Human Mental Workload, Peter A. Hancock and Najmedin Meshkati (Eds.). Advances in Psychology, Vol. 52. North-Holland, 139–183. https://doi.org/10.1016/S0166- 4115(08)62386-9 [10] Jeremy Howard and Sebastian Ruder. 2018. Universal Language Model Finetuning for Text Classification. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Association for Computational Linguistics, Melbourne, Australia, 328–339. https://doi.org/ 10.18653/v1/P18-1031 [11] Jiyou Jia. 2009. CSIEC: A computer assisted English learning chatbot based on textual knowledge and reasoning. Knowledge-Based Systems 22, 4 (2009), 249–255. https://doi.org/10.1016/j.knosys.2008.09.001 Artificial Intelligence (AI) in Blended Learning. [12] Ellen Jiang, Kristen Olson, Edwin Toh, Alejandra Molina, Aaron Donsbach, Michael Terry, and Carrie J Cai. 2022. PromptMaker: Prompt-Based Prototyping with Large Language Models. In Extended Abstracts of the 2022 CHI Conference on Human Factors in Computing Systems (New Orleans, LA, USA) (CHI EA ’22). Association for Computing Machinery, New York, NY, USA, Article 35, 8 pages. https://doi.org/10.1145/3491101.3503564 [13] Ellen Jiang, Edwin Toh, Alejandra Molina, Aaron Donsbach, Carrie J Cai, and Michael Terry. 2021. GenLine and GenForm: Two Tools for Interacting with Generative Language Models in a Code Editor. In Adjunct Proceedings of the 34th Annual ACM Symposium on User Interface Software and Technology (Virtual Event, USA) (UIST ’21 Adjunct). Association for Computing Machinery, New York, NY, USA, 145–147. https://doi.org/10.1145/3474349.3480209 [14] Ellen Jiang, Edwin Toh, Alejandra Molina, Kristen Olson, Claire Kayacik, Aaron Donsbach, Carrie J Cai, and Michael Terry. 2022. Discovering the Syntax and Strategies of Natural Language Programming with Generative Language Models. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems (New Orleans, LA, USA) (CHI ’22). Association for Computing Machinery, New York, NY, USA, Article 386, 19 pages. https://doi.org/10.1145/3491102.3501870 [15] Daniel Kahneman. 2011. Thinking, fast and slow. macmillan. [16] Tae Soo Kim, Yoonjoo Lee, Minsuk Chang, and Juho Kim. 2023. Cells, Generators, and Lenses: Design Framework for Object-Oriented Interaction with Large Language Models. In Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology. ACM. 1–18. [17] Johannes Kunkel, Benedikt Loepp, and Jürgen Ziegler. 2017. A 3D Item Space Visualization for Presenting and Manipulating User Preferences in Collaborative Filtering. In Proceedings of the 22nd International Conference on Intelligent User Interfaces (Limassol, Cyprus) (IUI ’17). Association for Computing Machinery, New York, NY, USA, 3–15. https://doi.org/10.1145/3025171.3025189 [18] Peter Lee, Sebastien Bubeck, and Joseph Petro. 2023. Benefits, Limits, and Risks of GPT-4 as an AI Chatbot for Medicine. New England Journal of Medicine 388, 13 (2023), 1233–1239. https://doi.org/10.1056/NEJMsr2214184 arXiv:https://doi.org/10.1056/NEJMsr2214184 PMID: 36988602. [19] Brian Lester, Rami Al-Rfou, and Noah Constant. 2021. The Power of Scale for Parameter-Efficient Prompt Tuning. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, Online and Punta Cana, Dominican Republic, 3045–3059. https: //doi.org/10.18653/v1/2021.emnlp-main.243 [20] Michael Xieyang Liu, Advait Sarkar, Carina Negreanu, Benjamin Zorn, Jack Williams, Neil Toronto, and Andrew D. Gordon. 2023. “What It Wants Me To Say”: Bridging the Abstraction Gap Between End-User Programmers and CodeGenerating Large Language Models. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (Hamburg, Germany) (CHI ’23). Association for Computing Machinery, New York, NY, USA, Article 598, 31 pages. https: //doi.org/10.1145/3544548.3580817 [21] Vivian Liu, Han Qiao, and Lydia Chilton. 2022. Opal: Multimodal Image Generation for News Illustration. In Proceedings of the 35th Annual ACM Symposium on User Interface Software and Technology (Bend, OR, USA) (UIST ’22). Association for Computing Machinery, New York, NY, USA, Article 73, 17 pages. https://doi.org/10.1145/3526113.3545621 [22] Vivian Liu, Jo Vermeulen, George Fitzmaurice, and Justin Matejka. 2023. 3DALLE: Integrating Text-to-Image AI in 3D Design Workflows. In Proceedings of the 2023 ACM Designing Interactive Systems Conference (Pittsburgh, PA, USA) (DIS ’23). Association for Computing Machinery, New York, NY, USA, 1955–1977. https://doi.org/10.1145/3563657.3596098 [23] Aditi Mishra, Utkarsh Soni, Anjana Arunkumar, Jinbin Huang, Bum Chul Kwon, and Chris Bryan. 2023. PromptAid: Prompt Exploration, Perturbation, Testing and Iteration using Visual Analytics for Large Language Models. arXiv:2304.01964 [cs.HC] [24] Joon Sung Park, Joseph C. O’Brien, Carrie J. Cai, Meredith Ringel Morris, Percy Liang, and Michael S. Bernstein. 2023. Generative Agents: Interactive Simulacra of Human Behavior. arXiv:2304.03442 [cs.HC] [25] Savvas Petridis, Nediyana Daskalova, Sarah Mennicken, Samuel F Way, Paul Lamere, and Jennifer Thom. 2022. TastePaths: Enabling Deeper Exploration and Understanding of Personal Preferences in Recommender Systems. In 27th International Conference on Intelligent User Interfaces (Helsinki, Finland) (IUI ’22). Association for Computing Machinery, New York, NY, USA, 120–133. https: //doi.org/10.1145/3490099.3511156 [26] Savvas Petridis, Nicholas Diakopoulos, Kevin Crowston, Mark Hansen, Keren Henderson, Stan Jastrzebski, Jeffrey V Nickerson, and Lydia B Chilton. 2023. AngleKindling: Supporting Journalistic Angle Ideation with Large Language Models. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (Hamburg, Germany) (CHI ’23). Association for Computing Machinery, New York, NY, USA, Article 225, 16 pages. https://doi.org/10.1145/3544548. 3580907 [27] Reid Pryzant, Dan Iter, Jerry Li, Yin Tat Lee, Chenguang Zhu, and Michael Zeng. 2023. Automatic Prompt Optimization with "Gradient Descent" and Beam Search. arXiv:2305.03495 [cs.CL] [28] Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. 2019. Language models are unsupervised multitask learners. OpenAI blog 1, 8 (2019), 9. [29] Kiran Ramesh, Surya Ravishankaran, Abhishek Joshi, and K. Chandrasekaran. 2017. A Survey of Design Techniques for Conversational Agents. In Information, Communication and Computing Technology, Saroj Kaushik, Daya Gupta, Latika Kharb, and Deepak Chahal (Eds.). Springer Singapore, Singapore, 336–350. [30] Emily Reif, Minsuk Kahng, and Savvas Petridis. 2023. Visualizing Linguistic Diversity of Text Datasets Synthesized by Large Language Models. arXiv:2305.11364 [cs.CL] [31] Laria Reynolds and Kyle McDonell. 2021. Prompt Programming for Large Language Models: Beyond the Few-Shot Paradigm. In Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems (Yokohama, Japan) (CHI EA ’21). Association for Computing Machinery, New York, NY, USA, Article 314, 7 pages. https://doi.org/10.1145/3411763.3451760 [32] Taylor Shin, Yasaman Razeghi, Robert L. Logan IV, Eric Wallace, and Sameer Singh. 2020. AutoPrompt: Eliciting Knowledge from Language Models with Automatically Generated Prompts. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). Association for Computational Linguistics, Online, 4222–4235. https://doi.org/10.18653/v1/2020. emnlp-main.346 [33] Hendrik Strobelt, Albert Webson, Victor Sanh, Benjamin Hoover, Johanna Beyer, Hanspeter Pfister, and Alexander M. Rush. 2023. Interactive and Visual Prompt Engineering for Ad-hoc Task Adaptation with Large Language Models. IEEE Transactions on Visualization and Computer Graphics 29, 1 (2023), 1146–1156. https://doi.org/10.1109/TVCG.2022.3209479 [34] Hongjin Su, Jungo Kasai, Chen Henry Wu, Weijia Shi, Tianlu Wang, Jiayi Xin, Rui Zhang, Mari Ostendorf, Luke Zettlemoyer, Noah A. Smith, and Tao Yu. 2022. Selective Annotation Makes Language Models Better Few-Shot Learners. arXiv:2209.01975 [cs.CL] [35] Gust Verbruggen, Vu Le, and Sumit Gulwani. 2021. Semantic Programming by Example with Pre-Trained Models. Proc. ACM Program. Lang. 5, OOPSLA, Article 100 (oct 2021), 25 pages. https://doi.org/10.1145/3485477 [36] Sitong Wang, Savvas Petridis, Taeahn Kwon, Xiaojuan Ma, and Lydia B Chilton. 2023. PopBlends: Strategies for Conceptual Blending with Large Language Models. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (Hamburg, Germany) (CHI ’23). Association for Computing Machinery, New York, NY, USA, Article 435, 19 pages. https://doi.org/10.1145/3544548. [37] Yunlong Wang, Shuyuan Shen, and Brian Y Lim. 2023. RePrompt: Automatic Prompt Editing to Refine AI-Generative Art Towards Precise Expressions. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (Hamburg, Germany) (CHI ’23). Association for Computing Machinery, New York, NY, USA, Article 22, 29 pages. https://doi.org/10.1145/3544548.3581402 [38] Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, brian ichter, Fei Xia, Ed Chi, Quoc V Le, and Denny Zhou. 2022. Chain-of-Thought Prompting Elicits Reasoning in Large Language Models. In Advances in Neural Information Processing Systems, S. Koyejo, S. Mohamed, A. Agarwal, D. Belgrave, K. Cho, and A. Oh (Eds.), Vol. 35. Curran Associates, Inc., 24824–24837. https://proceedings.neurips.cc/paper_files/paper/2022/file/ 9d5609613524ecf4f15af0f7b31abca4-Paper-Conference.pdf [39] Sherry Wu, Hua Shen, Daniel S Weld, Jeffrey Heer, and Marco Tulio Ribeiro. 2023. ScatterShot: Interactive In-Context Example Curation for Text Transformation. In Proceedings of the 28th International Conference on Intelligent User Interfaces (Sydney, NSW, Australia) (IUI ’23). Association for Computing Machinery, New York, NY, USA, 353–367. https://doi.org/10.1145/3581641.3584059 [40] Tongshuang Wu, Ellen Jiang, Aaron Donsbach, Jeff Gray, Alejandra Molina, Michael Terry, and Carrie J Cai. 2022. PromptChainer: Chaining Large Language Model Prompts through Visual Programming. In Extended Abstracts of the 2022 CHI Conference on Human Factors in Computing Systems (New Orleans, LA, USA) (CHI EA ’22). Association for Computing Machinery, New York, NY, USA, Article 359, 10 pages. https://doi.org/10.1145/3491101.3519729 [41] Tongshuang Wu, Michael Terry, and Carrie Jun Cai. 2022. AI Chains: Transparent and Controllable Human-AI Interaction by Chaining Large Language Model Prompts. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems (New Orleans, LA, USA) (CHI ’22). Association for Computing Machinery, New York, NY, USA, Article 385, 22 pages. https: //doi.org/10.1145/3491102.3517582 [42] Yu Wu, Wei Wu, Chen Xing, Ming Zhou, and Zhoujun Li. 2017. Sequential Matching Network: A New Architecture for Multi-turn Response Selection in RetrievalBased Chatbots. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Association for Computational Linguistics, Vancouver, Canada, 496–505. https://doi.org/10.18653/v1/P17-1046 [43] Anbang Xu, Zhe Liu, Yufan Guo, Vibha Sinha, and Rama Akkiraju. 2017. A New Chatbot for Customer Service on Social Media. In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems (Denver, Colorado, USA) (CHI ’17). Association for Computing Machinery, New York, NY, USA, 3506–3510. https://doi.org/10.1145/3025453.3025496 [44] Ann Yuan, Andy Coenen, Emily Reif, and Daphne Ippolito. 2022. Wordcraft: Story Writing With Large Language Models. In 27th International Conference on Intelligent User Interfaces (Helsinki, Finland) (IUI ’22). Association for Computing Machinery, New York, NY, USA, 841–852. https://doi.org/10.1145/3490099.3511105 [45] J.D. Zamfirescu-Pereira, Heather Wei, Amy Xiao, Kitty Gu, Grace Jung, Matthew G Lee, Bjoern Hartmann, and Qian Yang. 2023. Herding AI Cats: Lessons from Designing a Chatbot by Prompting GPT-3. In Proceedings of the 2023 ACM Designing Interactive Systems Conference (Pittsburgh, PA, USA) (DIS ’23). Association for Computing Machinery, New York, NY, USA, 2206–2220. https://doi.org/10.1145/3563657.3596138 [46] J.D. Zamfirescu-Pereira, Richmond Y. Wong, Bjoern Hartmann, and Qian Yang. 2023. Why Johnny Can’t Prompt: How Non-AI Experts Try (and Fail) to Design LLM Prompts. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (Hamburg, Germany) (CHI ’23). Association for Computing Machinery, New York, NY, USA, Article 437, 21 pages. https://doi.org/10.1145/ 3544548.3581388 [47] Tianyi Zhang, London Lowmanstone, Xinyu Wang, and Elena L. Glassman. 2020. Interactive Program Synthesis by Augmented Examples. In Proceedings of the 33rd Annual ACM Symposium on User Interface Software and Technology (Virtual Event, USA) (UIST ’20). Association for Computing Machinery, New York, NY, USA, 627–648. https://doi.org/10.1145/3379337.3415900 [48] Hao Zhou, Minlie Huang, Tianyang Zhang, Xiaoyan Zhu, and Bing Liu. 2018. Emotional Chatting Machine: Emotional Conversation Generation with Internal and External Memory. arXiv:1704.01074 [cs.CL] This paper is under CC 4.0 license. available on arxiv

Interactively Critiquing LLMs by Converting Feedback into Principle— Conclusion and References

A Conversation with Ilnar Shafigullin Today, we are privileged to converse with Ilnar Shafigullin, a notable methodology specialist with an expansive expertise in applied mathematics, data science, and AI/ML methodologies and solutions, particularly within the innovative sphere of EdTech platforms. Ilnar, with a substantial 12-year history in university teaching and a deep-seated passion for educational methodologies, has been an influential figure in the field. In recent years, he has been steering the methodological direction at Geekbrains, a preeminent entity in the realm of EdTech in CIS and a luminary in IT education, holding the #1 position in the CIS. With a staggering 70K+ students enrolled in programs spanning 1-3 years and a user base of over 6M+, Geekbrains stands as a monumental presence in the educational technology sector. As its Methodology Director, Ilnar has significantly impacted the educational narrative, notably through his pioneering "Developer" educational product. As we unravel the intricacies of modern educational trajectories, explore the prevailing challenges, and contemplate the anticipated future of learning, let’s embark on this explorative discourse. The past decade has witnessed education metamorphosing in unprecedented ways, transitioning online and further morphing with the advent of generative networks. How do you perceive the contemporary educational market? Could you shed light on the essence of education, its constituents, and its multifaceted forms? Alexander Volchek: To truly grasp the essence of education, we need to appreciate its multifaceted nature, influenced by a myriad of parameters. Education can manifest as online, offline, or a hybrid blend; it can be synchronous or asynchronous; orchestrated for individuals or groups, and so forth. The permutations are vast. For instance, traditional schooling predominantly hinges on offline, synchronous group sessions. Contrast this with online foreign language tutoring, which leans towards an individual, synchronous mode. And mind you, these are just the tip of the iceberg. The myriad combinations underscore the expansive canvas that is "education." Ilnar Shafigullin: It's also pivotal to understand that there isn't a one-size-fits-all formula. The choice of educational approach hinges heavily on the objectives in play, whether from the student's lens, the EdTech entity's perspective, or the educator's vantage point. Zooming into my most recent endeavors over the last few years, I've been deeply entrenched in large-scale, blended synchronous-asynchronous online education tailored for the IT realm. As it stands, programs that I've helmed have tens of thousands of students onboard, with durations spanning from 6 to a staggering 36 months. Let's delve deeper into the vast expanse of large-scale online education. Could you shed light on its intricacies and the prevailing trends in recent times? Alexander Volchek: The landscape is dotted with a myriad of entities, each charting their unique course. Yet, when we distill their essence from a business lens, the overarching ambition for EdTech ventures invariably circles back to maximizing profitability. Navigating this quest, several pivotal educational challenges emerge: Ilnar Shafigullin: Scalable Education: The imperative is not just to cater to the masses but to ensure scalability without diluting educational quality. Cost Efficiency: Every student's operational expenditure needs to be streamlined to the bare minimum. High Completion and Referral Rates: The gold standard is not just student retention but ensuring they realize their set educational aspirations. The ripple effect? A satisfied student body that becomes ambassadors of your product. The market's diverse players adopt varied strategies to tackle these challenges. A popular approach hinges on asynchronous, online, and solo learning paradigms. Here, the learning ecosystem orbits around students autonomously navigating pre-curated materials and undertaking assignments. These tasks either undergo automated assessments or are vetted by educators. Whether it's text-based content or video lectures, this model stands out in scalability and cost-effectiveness. However, this model grapples with the challenge of ensuring tangible educational outcomes. Given the absence of real-time interactions with educators or peers, the onus of success squarely rests on the student's shoulders. Their intrinsic motivation and tenacity become the linchpins of their educational journey. My professional odyssey resonated with an establishment that championed synchronous, online, group-centric learning. Here, each session transforms into a virtual conclave of students and a subject maestro, unraveling new topics and collaboratively tackling tasks. The allure of this model lies in its heightened student engagement, fostered by collective learning and immediate access to experts. But it's not without its challenges, especially when scaling becomes the ambition. For context, teaching a cohort of 50 students might necessitate one expert. But scaling this to 5,000 students implies rallying an army of 100 experts, each mirroring the same expertise level. The inevitable fallout? Skyrocketing operational costs. Fascinating insights. Confronted with such intricate challenges, how did you navigate this maze to ensure scalability without compromising quality? Interviewer: It's a multifaceted conundrum, no doubt. Our leadership was unequivocal in its vision: preserve the essence of synchronous education while championing scalability and fostering operational frugality. And as the numbers suggest, with over 30,000 students engaged with our product, we've indeed hit our stride. Ilnar Shafigullin: To unpack this, it's pivotal to demystify the learning process. What does it encapsulate? Fundamentally, it's about imparting a chunk of theoretical knowledge to students and subsequently ensuring they not only grasp it but can pragmatically apply it. If there's a disconnect, the onus is on us to bridge that gap. In an undiluted synchronous paradigm, both these facets resist effortless scalability. Especially the preliminary phase, the dissemination of theoretical knowledge, is a veritable minefield. Hypothetically, let's say our arsenal comprises 100 experts tasked with enlightening our students. The expectation is uniformity - each educator should exude consistent expertise, radiate enthusiasm, and be adept at student interactions. Each of these sessions carries a price tag, and an expert's caliber directly inflates the cost. Yet, stripped to its core, they're essentially echoing the same content. The logical pivot? Transitioning such sessions into the asynchronous realm. Imagine harnessing the prowess of a singular, top-tier expert, recording it in a state-of-the-art studio. Every student would then access this meticulously crafted lecture, potentially reaping more value than from a live session helmed by a mid-tier expert. The ensuing challenge, ensuring the assimilation and application of imparted knowledge, still merits synchronous engagement. But with a twist. We could curate a blueprint for these sessions, laden with well-defined tasks, probing questions, and a roadmap for student interactions. Supplement this with exhaustive solutions and preemptive answers, and even a novice instructor can steer the session to success. This recalibration not only mitigates our dependence on high-caliber experts but also catapults our product's scalability and, importantly, accentuates the educational experience. Still, this is but a glimpse into the myriad adjustments essential to uphold the sanctity of synchronous education. Indeed, the details you've shared are eye-opening. Can you shed light on any other intricacies associated with synchronous education? Interviewer: Certainly. When you cast your gaze upon the titans of the industry—entities that don't confine themselves to niche domains but endeavor to enlighten students across an expansive spectrum of disciplines—a glaring challenge presents itself: maintaining the gold standard of educational content and orchestrating the formation of cohorts. Given our commitment to synchronous learning, operational overheads associated with conducting classes are significant. Consequently, the initiation of any teaching cohort mandates a specific student count to ensure financial viability. Ilnar Shafigullin: Both these challenges find their solution in a radical reshaping of learning trajectories. I've pioneered a tree-structured learning framework within our entity. This strategy emphasizes front-loading the foundational components of the learning journey—those segments that demand infrequent updates. The more volatile, frequently updated segments are positioned towards the trajectory's culmination. This stratagem enabled us to unify certain segments of diverse learning pathways. All students embark on a unified journey for the initial two months. Subsequently, a career alignment phase bifurcates them into specialized paths, and another seven months down the line, they diverge based on their chosen tech stacks. This approach not only optimized our unit economics in an unprecedented manner but also guaranteed that our student cohorts reached a critical mass, facilitating bi-monthly commencement of courses. In essence, we've managed to retain the essence of synchronous learning, scale it to accommodate tens of thousands, and sculpt an unparalleled economic model for our segment. This has been vividly demonstrated through the Developer program, a project I've passionately been building for the past 2 years, where all these methodologies and principles have not only been implemented but have also proven their efficacy and impact in real-world educational contexts. That's truly commendable! Casting your gaze forward, how do you envision the trajectory of education and the EdTech landscape, especially with the dawn of transformative technologies like generative networks? Interviewer: Generative networks are potent catalysts poised to supercharge the educational realm. Traditional education invariably grapples with the constraints of human bandwidth— the responsiveness of an expert, the turnaround time for assignment evaluations, or elucidations of intricate lecture points. And let's not forget, experts come with a premium. Generative networks, in this context, can potentially revolutionize the triad of challenges I previously outlined: scalability, operational expenditure, and realizing educational outcomes. Ilnar Shafigullin: While integrating these nascent technologies does pose certain challenges, primarily due to their novelty and the limited fraternity with access to such tech, the trajectory is clear. Generative networks are poised to make their indelible mark on the educational tapestry. Virtually every facet of the learning ecosystem stands to gain from these networks, signaling an exhilarating future. A heartfelt thank you to Ilnar for sharing such a comprehensive and enlightening narrative about contemporary IT education. It's truly been an insightful journey into understanding the intricate, multifaceted nature of the educational process. If any of our readers are contemplating venturing into the education sector, it's crucial to recognize and appreciate its multifaceted nature. Ilnar has gracefully unfolded just a glimpse of a few parameters, illustrating an immensely varied approach towards learning and educational methodologies, and there is certainly a profound depth to explore further in this narrative. Moreover, the emphasis on integrating AI into educational practices is not to be dismissed as mere trend-following; it's a tangible leap towards the future. The imminent years are poised to witness transformative shifts in education, significantly shaped by the introduction and integration of generative networks. It's a field ripe for exploration, experimentation, and practical application of these networks. Thus, delve, explore, and experiment with generative networks as they undeniably play a pivotal role in scripting the future trajectory of education. Your journey into understanding and perhaps shaping the future of education starts with recognizing these emerging, transformative trends.

Navigating the Future of EdTech: Synchronous Learning and the Power of Generative Networks 

11 Real-World Applications of Machine Learning in Ecommerce 

11 aplicaciones del mundo real del aprendizaje automático en el comercio electrónico

11 applications réelles de l'apprentissage automatique dans le commerce électronique

ईकॉमर्स में मशीन लर्निंग के 11 वास्तविक-विश्व अनुप्रयोग

電子商取引における機械学習の 11 の実世界の応用

11 aplicações do mundo real de aprendizado de máquina no comércio eletrônico

11 ứng dụng thực tế của Machine Learning trong thương mại điện tử

机器学习在电子商务中的 11 个实际应用

Hello everyone! We are a small team that has been developing a 2D strategy game called - a medieval kingdom simulator - for several years now. Norland This is our second article about the 2D of various effects in our game. You can read the previous article . Just to remind you, the game is a 2D one and is being developed on the Game Maker Studio 2 engine. rendering here Today, we will be discussing four elements - water, fire, earth, and air. What graphical effects do they encompass? Let's find out. Water Currently, water doesn't have a profound impact on the in Norland - you can't build structures on it, and that's about it. You can't catch fish or construct ships yet. Nevertheless, it still needs to be rendered somehow. gameplay It all started with the development of a mockup (a fake screenshot of the game drawn in a graphics editor). As the initial water reference, I used a screenshot from the game Graveyard Keeper, adjusted to our color palette. Artists created the tiles for the shoreline for us, but they struggled with the actual water surface. So, I had to come up with its appearance myself. The coastline tiles were drawn by our artists, but they had some difficulties with creating realistic water surfaces. Therefore, I had to come up with the appearance of the water myself. I had a rough understanding of how to create a shader for water - a seamless texture with "caustics" is overlaid on water tiles, and waves and ripples are simulated by moving the UV coordinates of the water surface over a special noise texture (effect commonly known as Displacement). I found the caustics on , but I don't remember where I got the texture for shifting the UV coordinates - they are found online under various names (e.g., bump mapping water texture). OpenGameArt The effect is simple - at each point, I take the r and b components of the Displacement texture color, which represent the shift value, and add it to the texture coordinates of the caustics. Additionally, the Displacement texture is looped and constantly shifted in a certain direction (which corresponds to the direction of water ripples). I recreated this shader in , and if you're interested, you can view its implementation and experiment with your own parameter values and textures. The project can be downloaded . SHADERed here Fire Initially, to create the effect of burning houses, only particles were used. However, I quickly encountered a problem: to achieve a dense and vibrant fire effect, a significant number of particles were required. This, in turn, negatively affected the game's performance (Game Maker Studio 2.3 is not the most performant engine), especially when a fire spread across half of the city. Moreover, I wasn't satisfied with the resulting effect - the fire seemed insubstantial and didn't blend well with the environment. You can see the results of this work in the game's announcement trailer. https://youtu.be/rUvyOVP0HTo?t=55&embedable=true So, to avoid game lag, it was necessary to reduce the number of particles. However, even with the initial particle count, the effect still appeared insufficiently dense. Therefore, a different approach was needed. In the end, I came up with a solution that involved a small number of particles and a special fire shader applied to the building texture, creating the desired density of the effect. The effect consists of several components: Creating a fire texture from Perlin noise, overlaying this texture on the building, and animating the result (simple upward movement of the fire); Darkening the building texture and subsequently "dissolving" it (effect commonly known as Dissolve) to fully destroy the building engulfed in fire; Several types of particles - smoke, scattered sparks from the fire, and an older fire effect that was used on a much smaller scale than before. I also this effect to SHADERed. transferred Earth If you simply apply a tiled texture to the surface, it doesn't matter how good the texture is - as the camera moves away, it will look unpleasant. To achieve an acceptable result, it is necessary to somehow break the obvious repetitiveness of the pattern. Initially, I had the idea to generate various-shaped and colored grass clumps and then randomly scatter them across the entire map. I was satisfied with this approach - it lasted for almost a year, but as the game is getting closer to release, it was time to revisit this task for polishing. The downside of using clumps became evident - to ensure even coverage of a map sized 31,500 x 22,500, a large number of decals would be required, and they would inevitably intersect with each other, creating overdraw (excessive rendering of pixels that won't be visible on the screen due to object overlap). And it turned out that there might not be enough of them. Moreover, the shape of these clumps is predetermined, so creating new ones would require manual work in the editor. Therefore, I came up with another solution - generating a diverse surface pattern directly in the (GLSL Fragment). shader The code for this shader is relatively simple - a tiled Perlin noise texture is fed into it, and the function is used to cut off areas that are darker than a threshold value. smoothstep I perform this operation twice, but the second time, the noise texture is scaled and offset differently to avoid repeating patterns. Then, the spots are colored with designated colors and blended together based on alpha, and finally, a tiled grass texture is overlaid. In addition to the texture itself, various decals such as grass clumps, rocks, and dirt spots are placed on top of the grass. However, they are required in much smaller quantities than before. Here's the result before and after the effect. Air Actually, this section is not specifically about air but rather about color grading. However, I needed to align it with the overall concept of natural elements in the game ;) LUT (Look-Up Table) is a technique that is well-known in the photography and videography communities. Its purpose is the same everywhere – to colorize an image in a specific way. A LUT is essentially a texture (usually 512x512 pixels) that encodes the entire RGB color range. Imagine it as a cube divided into layers. The texture may look something like this: The idea is that each RGB color corresponds to a specific pixel in this texture (the RGB components of the color are used as coordinates in the texture). For example, for the color red (#ff0000), it would be the top-right pixel in the first square. The job of a color grading shader is to change the color of each pixel in the image to the color corresponding to its pixel in the LUT texture. The only thing you need to do manually is prepare these LUT textures, but it's a fairly straightforward process. For example, let's say we want to create the effect of a sunset with a shift toward red tones. We take a screenshot from the game, load it into a photo editor, and start playing with color parameters using various tools like Hue/Saturation, Brightness, Contrast, etc. Personally, I would focus on shifting the Hue toward red and increasing saturation, but feel free to let your imagination run wild. You could, for example, saturate only the red color and make everything else gray (hello, Sin City). Next, you apply the same color adjustments to a neutral LUT texture. The result would look like this: Now, if you of our game through the color grading shader with this texture, the game will look exactly as the image was adjusted in the photo editor. pass the render In Norland, we use a set of five different LUT textures, each corresponding to a specific time of day – morning, day, evening, sunset, night – as well as a special texture for artificial lighting. At any given time, two textures are active and blended together to achieve a smooth transition in color grading. For example, if morning starts at 6 am and day starts at 12 pm, at 8 am the color grading will be a mix of 33% morning LUT and 66% day LUT. There are many other effects in the game (artificial lighting, rain, blood, equipment damage, etc.) of varying complexity, but there's still much work to be done (I'm currently brainstorming how to create snow and winter). But that's a story for another time. I hope you found it interesting! Thank you! Also published here.

The Four Elements and Their Graphic Effects: Exploring Shaders in a 2D Game

Star-Clouds, Star-Clusters, and Star-Streams

The “Backendification” of Frontend Development

design refers to the process of creating interfaces within the software or computerized devices, focusing on appearance and style. Here, designers mainly intend to create designs that users would find easy, intuitive, and pleasurable to use. User interface (UI) Typically, UI design refers to the graphical user interface (GUI), but there are others, including the voice-controlled interfaces. UI is an important part of the design process as it plays an important role in the customer acquisition process, customer retention, lowering support costs, increasing productivity and reducing development time. (UI is often confused with user experience — but both are two different aspects that are closely related and aim at driving user satisfaction.) You risk losing your customers if they don’t find your app or site’s UI appealing at first glance because this is going to be the first thing that they notice when they reach your site. According to certain researches, around 26% of apps are not used more than once owing to the poor UI. To ensure that people engage with your UI, it is important to implement the latest trends. With the latest UI trends, you are promising an improved experience and enhanced usability, which contributes to building trust. After extensive research, we have come up with some of the UI trends that would pave the way to great user experience (UX) in the year 2019: 1. Typography and Storytelling Font style has an important role in driving user attention to content.Storytelling that revolves around a digital experience would be a gamechanger in 2019. Typography has a role in influencing the visual hierarchy of the website. This is one of the reasons why it becomes a vital aspect of UI — which, in turn, enhances user experience. Though text style will help in capturing user attention, without a powerful content, we are far from offering a rich user experience. Typography and copywriting work hand in hand — style attracts user attention, while great content engages users with the brand — making them a part of the tale that the brand narrates. The main intent is to prevent users from skimming through the content and leaving the website, without taking home a value-added solution to their concerns. While the style captures the user’s attention, the narrative engages users with the brand because it makes them a part of the story. This, in turn, keeps users from simply scanning through the content, but deriving a value from the narrative. 2. Micro-animations and custom illustrations Micro-animations have been around for a while — and will continue in the year 2019. One of the main advantages of micro-animations is that these help visitors to distinguish elements on the screen, thereby enhancing the interaction with users. In the previous year, custom illustrations, regardless of whether these were hand-drawn or digital, had gained popularity and will continue to be used widely in 2019. The vast asymmetries, free forms, and unaligned designs, which are part of the micro-animations ensure that the websites appear unique and stand out in the crowd, while also creating a user-friendly environment — promising a great user experience. The illustrations when combined with complex motions — breathe life into the pages. And this blend of motion with illustrations captures user attention — providing the visitor with an understanding of what the brand represents. 3. Colors, Gradients and Overlapping Effect The overlapping effect is expected to become popular over the upcoming days. This effect comprises color, text or image. With overlapping effects, you can ensure that there is sufficient space amid the different elements to attract user attention. Gradients disappeared at some point in time a few years back — but made a comeback in 2018 and is expected to be a trend in 2019. Gradients are everywhere — within illustrations, text and UI elements. With gradients, realism and depth get added to the platform, which offers a real-life experience as the users move through the site or app. 4. Immersive, full-screen videos Unlike any other elements, such as infographics and images, visuals have a deeper impact on driving user attention. But the latest UI trend is to enlarge videos so that these fit into the entire screen space. This UI trend has been around for a while now — UI designers use it a means of storytelling and engagement — and it would be a smart idea to include these to enhance your UI, going forward. These are actually images in motion and bigger in size, and also capable of engaging the audience. When these videos are blended with bold and perfect typography, they are bound to leave an impact on the users — conveying the brand’s message. Amid the traditional photography that is static, people love to see more videos that are a feast to the eyes. Along with conventional imagery, rich, interactive and full-screen videos would be used as a means of storytelling and user engagement in 2019 — in fact, video clips are more appealing than watching static images. 5. Long scrolling Scrolling is always a part of the features that enhance UX and it will be there forever. But we are witnessing a lot of experimentations using scrolling to enhance UI design. This time scrolling would be more focused on providing users with the provision to read a long-form content without having to switch across pages. Recommended Reading: Top 10 Web Design Trends in 2019 Now, designers are using the long-scroll form to make room for long content. This helps the users to read through the information without having to navigate to other pages, but with a single scroll through the page that they are reading. This is the best approach to storytelling. Apart from storytelling, it helps in viewing a structured content in a linear way. 6. Cards Cards can be described as little rectangles that hold inclusive images and text. These elements serve as entry points to more detailed information. These are referred to as cards because these resemble real-world tangible cards in UIs. Cards were always there — in the form of sticky notes, business cards and what not — but now, we are using these more often in web and mobile apps. When implemented correctly, digital cards help in enhancing the UX aspect of apps. Some of the reasons for using cards are listed below: Organizing content — cards help to organize information well — which aids for scannability when the users read through. With cards, content can be divided into meaningful sections — which again makes it easier for users to find what they are looking for quickly. The ideal choice for varying screen sizes — since cards are infinitely manipulative, the varying screen size feature helps to create a single design that can be used across multiple devices, in order to ensure a consistent experience, no matter what device is being used. 7. Full-screen backgrounds One of the trends that were observed in the year 2018 was full-screen backgrounds and this is expected to be on the trend list in 2019. The main advantage of full-screen backgrounds is to gift the screens with a visual appeal to attract user attention. For websites with minimal content and less complex structure, full-screen images are ideally the best choice. These large backgrounds are not just images, but also animations, videos or illustrations can also be used as full-screen backgrounds. 8. Buttonless UI Buttons have always been a vital part of UI design — it will continue to be in the year 2019. It is expected that there would be more creative experiments on mobile user experience from UI interaction perspective, without using buttons. (Voice search, scrolling, and drag and drop are some of the examples of buttonless UI) The advantage of a buttonless UI is that it helps in saving a lot of space on the screen, which can be utilized for placing other elements, such as content. 9. High-quality photo content One more UI aspect that would make it to the trend list in 2019 is high-quality photo content. Excellent and high-quality images attract user attention. Realistic and clear visuals are magic when it comes to driving user attention. Stock photos were being used to attract user attention all this time. But rather than using images that are repetitive and that do not relate well with the content on the website, realistic photos can do the work. Of course, you can find great photos with Shutter Stock and other websites — but, when it comes to creating a human touch, only real images are ideal choices. This is specifically important when you are working on specific domains, such as fashion, real estate, travel and tourism, and restaurants. Let us create magic with real-life images — spend a little on photography and witness awesome results for yourself. Designing UIs UIs are the access point where the user interaction with designs happens. Designers main intention is to create interfaces that users will find highly efficient and usable. It is important to ensure that a smooth and seamless interaction is offered to the users through the UIs you design. We see newer trends being introduced in UI every year to enhance the UX. UI should not be just about the appeal, but usability, ease of use and functionality. It is important to implement UI trends because it would give your design a touch of the much-needed freshness and newness. We would like to hear from you. Please share your thoughts or suggestions in the comments box below.

BLEND-ED

blend-ed.com

-- employees

Since n.d.

COMPANY RANKING

BLEND-ED

EVERGREEN INDEX #2342

Representing this company? Click here to claim and customize this page!

Blend-ed WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!

BLEND-ED

blend-ed.com

-- employees

Since n.d.

COMPANY RANKING

BLEND-ED

EVERGREEN INDEX #2342

RELATED COMPANIESRANK

STORIES ABOUT BLEND-ED

Representing this company? Click here to claim and customize this page!

Blend-ed WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!