As seen on HYPEBEAST, CNBC, Highsnobiety, COMPLEX, NiceKicks, and more - HYP is the social auction house for exclusive releases

The Hypersonic Drone Market: How the Axis of Innovation Is Forming Around It

Lifelong Intelligence Beyond the Edge using Hyperdimensional Computing: Conclusion, and References

Authors:
(1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu);
(2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu);
(3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx);
(4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu);
(5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction 2 Related Work 3 Background on HDC 4 Problem Definition 5 LifeDH 6 Variants of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 9 DISCUSSIONS AND FUTURE WORKS Problem Scale. One limitation of LifeHD is the relative small problem scale (e.g., the image size of CIFAR-100 is restricted to 32x32) due to the essential difficulty of unsupervised lifelong learning problem, including single-pass non-iid data and no supervision. For the same reason, there remains a disparity in accuracy between unsupervised lifelong learning and fully supervised NNs, as substantiated by prior research [13, 54]. In order to scale LifeHD to more challenging applications such as self-driving vehicles, one possible direction is to leverage the pretrained foundation model as a frozen feature extractor in the HDnn framework, which we leave for future investigation. Hyperparameter Tuning. While we recognize that hyperparameters can influence the performance of LifeHD, such an issue is not exclusive to LifeHD, but has persistently been a challenge in machine learning research [7]. In LifeHD, the impact of hyperparameters can be mitigated through pre-deployment evaluation and component co-design. For example, encoding parameters such as 𝑄, 𝑃 can be tuned on similar health monitoring data sources prior to deployment. Meanwhile, the component of cluster HVs merging can increase LifeHD’s resiliency to the novelty detection threshold 𝛾, as a higher quantity of novel clusters can be merged in later stage of learning. Limitations of HDC. HDC serves as the fundamental core of LifeHD. While HDC shows promise with its notable lightweight design, it is burdened by several limitations that remain active areas of research. First, for complex datasets like audio and images, HDC requires a pretrained feature extractor (the HDnn encoding) which may not exist for certain applications. Moreover, akin to any other architecture, HDC vectors face capacity limitations determined by the dimension of HD space, encoding method, and potential noise levels in the input data [56]. Due to these factors, careful evaluation and sometimes manual feature engineering are required to successfully deploy HDC for new applications. Future Works. Although LifeHD focuses on single-device lifelong learning for classification tasks, the method can be extended for other types of tasks and learning settings, such as federated learning and reinforcement learning. We leave the investigation of these topics for future work. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Authors: Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction Abstract and 1. Introduction 2 Related Work 2 Related Work 3 Background on HDC 3 Background on HDC 4 Problem Definition 4 Problem Definition 5 LifeDH 5 LifeDH 6 Variants of LifeHD 6 Variants of LifeHD 7 Evaluation of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 10 Conclusion, Acknowledgments, and References 9 DISCUSSIONS AND FUTURE WORKS Problem Scale. One limitation of LifeHD is the relative small problem scale (e.g., the image size of CIFAR-100 is restricted to 32x32) due to the essential difficulty of unsupervised lifelong learning problem, including single-pass non-iid data and no supervision. For the same reason, there remains a disparity in accuracy between unsupervised lifelong learning and fully supervised NNs, as substantiated by prior research [13, 54]. In order to scale LifeHD to more challenging applications such as self-driving vehicles, one possible direction is to leverage the pretrained foundation model as a frozen feature extractor in the HDnn framework, which we leave for future investigation. Problem Scale. Hyperparameter Tuning. While we recognize that hyperparameters can influence the performance of LifeHD, such an issue is not exclusive to LifeHD, but has persistently been a challenge in machine learning research [7]. In LifeHD, the impact of hyperparameters can be mitigated through pre-deployment evaluation and component co-design. For example, encoding parameters such as 𝑄, 𝑃 can be tuned on similar health monitoring data sources prior to deployment. Meanwhile, the component of cluster HVs merging can increase LifeHD’s resiliency to the novelty detection threshold 𝛾, as a higher quantity of novel clusters can be merged in later stage of learning. Hyperparameter Tuning. Limitations of HDC. HDC serves as the fundamental core of LifeHD. While HDC shows promise with its notable lightweight design, it is burdened by several limitations that remain active areas of research. First, for complex datasets like audio and images, HDC requires a pretrained feature extractor (the HDnn encoding) which may not exist for certain applications. Moreover, akin to any other architecture, HDC vectors face capacity limitations determined by the dimension of HD space, encoding method, and potential noise levels in the input data [56]. Due to these factors, careful evaluation and sometimes manual feature engineering are required to successfully deploy HDC for new applications. Limitations of HDC. Future Works. Although LifeHD focuses on single-device lifelong learning for classification tasks, the method can be extended for other types of tasks and learning settings, such as federated learning and reinforcement learning. We leave the investigation of these topics for future work. Future Works. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. available on arxiv

Lifelong Intelligence Beyond the Edge using Hyperdimensional Computing: Discussions and Future Works

Authors:
(1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu);
(2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu);
(3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx);
(4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu);
(5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction 2 Related Work 3 Background on HDC 4 Problem Definition 5 LifeDH 6 Variants of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 8 EVALUATION OF LIFEHDsemi AND LIFEHDa In this section, we compare LifeHDsemi and LifeHDa, our proposed extensions from LifeHD, with existing designs that are similar. Performance of LifeHDsemi. To evaluate LifeHDsemi in a lowlabel scenario, we compare it with SemiHD [22], which is the stateof-the-art HDC method for semi-supervised learning. We adapt SemiHD [22] for single-pass settings, introducing a pseudolabel assignment threshold. When the cosine similarity of an unlabeled sample to the nearest class hypervector surpasses the threshold, we assign that class as its pseudolabel. The sample is then employed to update the class hypervector in SemiHD. We explore various threshold values and choose the optimal result for comparison. Fig. 15 (a) compares LifeHDsemi and SemiHD [22] on ESC-50 and CIFAR100 across various labeling ratios 𝑟 < 0.01. The advantages of LifeHDsemi are most prominent when labels are limited, the weakly supervised scenario is LifeHDsemi’s major focus. LifeHDsemi improves ACC by up to 10.25% and 3.6% on ESC-50 and CIFAR-100 respectively. This outcome arises from the unsupervised nature of LifeHD, allowing it to autonomously organize prominent cluster HVs, especially when all samples from a class lack labels. As the labeling ratio increases, LifeHDsemi’s advantage over SemiHD diminishes, because more labels bolster SemiHD’s performance. Performance of LifeHDa. LifeHDa provides an interface to trade minimal performance loss for efficiency gains, by adaptively pruning out the insignificant dimensions. We compare LifeHDa with previous HDC works employing a fixed mask throughout training [25], and the results are presented in Fig. 15 (b) for CIFAR100, including ACC and training latency per batch on RPi 4B. Fixed masks negatively impact HDC learning, especially with smaller dimensions. Such masks fail to adapt to new hypervectors in classincremental streams, where less significant dimensions may become crucial later in training. LifeHDa addresses this issue by adjusting the mask upon novelty detection, leading to a degradation of only 0.71% in ACC and 4.5x efficiency gain compared to the complete LifeHD, using only 20% of the full HD dimension of LifeHD. The overhead of adaptively adjusting the mask is negligible when novelty detection occurs infrequently. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Authors: Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction Abstract and 1. Introduction 2 Related Work 2 Related Work 3 Background on HDC 3 Background on HDC 4 Problem Definition 4 Problem Definition 5 LifeDH 5 LifeDH 6 Variants of LifeHD 6 Variants of LifeHD 7 Evaluation of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 10 Conclusion, Acknowledgments, and References 8 EVALUATION OF LIFEHDsemi AND LIFEHDa In this section, we compare LifeHDsemi and LifeHDa, our proposed extensions from LifeHD, with existing designs that are similar. Performance of LifeHDsemi. To evaluate LifeHDsemi in a lowlabel scenario, we compare it with SemiHD [22], which is the stateof-the-art HDC method for semi-supervised learning. We adapt SemiHD [22] for single-pass settings, introducing a pseudolabel assignment threshold. When the cosine similarity of an unlabeled sample to the nearest class hypervector surpasses the threshold, we assign that class as its pseudolabel. The sample is then employed to update the class hypervector in SemiHD. We explore various threshold values and choose the optimal result for comparison. Fig. 15 (a) compares LifeHDsemi and SemiHD [22] on ESC-50 and CIFAR100 across various labeling ratios 𝑟 < 0.01. The advantages of LifeHDsemi are most prominent when labels are limited, the weakly supervised scenario is LifeHDsemi’s major focus. LifeHDsemi improves ACC by up to 10.25% and 3.6% on ESC-50 and CIFAR-100 respectively. This outcome arises from the unsupervised nature of LifeHD, allowing it to autonomously organize prominent cluster HVs, especially when all samples from a class lack labels. As the labeling ratio increases, LifeHDsemi’s advantage over SemiHD diminishes, because more labels bolster SemiHD’s performance. Performance of LifeHDsemi. Performance of LifeHDa. LifeHDa provides an interface to trade minimal performance loss for efficiency gains, by adaptively pruning out the insignificant dimensions. We compare LifeHDa with previous HDC works employing a fixed mask throughout training [25], and the results are presented in Fig. 15 (b) for CIFAR100, including ACC and training latency per batch on RPi 4B. Fixed masks negatively impact HDC learning, especially with smaller dimensions. Such masks fail to adapt to new hypervectors in classincremental streams, where less significant dimensions may become crucial later in training. LifeHDa addresses this issue by adjusting the mask upon novelty detection, leading to a degradation of only 0.71% in ACC and 4.5x efficiency gain compared to the complete LifeHD, using only 20% of the full HD dimension of LifeHD. The overhead of adaptively adjusting the mask is negligible when novelty detection occurs infrequently. Performance of LifeHDa. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. available on arxiv

Lifelong Intelligence Beyond the Edge using Hyperdimensional Computing: Evaluation of LifeHDsemi & a

Lifelong Intelligence Beyond the Edge using Hyperdimensional Computing: Evaluation of LifeHD

Authors:
(1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu);
(2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu);
(3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx);
(4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu);
(5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction 2 Related Work 3 Background on HDC 4 Problem Definition 5 LifeDH 6 Variants of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 6 VARIANTS OF LIFEHD While LifeHD is designed to cater to general IoT applications with streaming input and without supervision, real-world scenarios may vary. Some scenarios might have a few labeled samples in addition to the unlabeled stream, while others may require operation within strict power constraints. LifeHD offers extensibility to address these diverse needs. In this section, we introduce two software-based extensions: LifeHDsemi, which adds a separate processing path to manage labeled samples, and LifeHDa, which adaptively prunes the HDC model using masking to handle low-power scenarios. 6.1 LifeHDsemi 6.2 LifeHDa While HDC computation is typically lightweight, there may be instances of energy scarcity (e.g., when powered by a solar panel) that call for a balance between accuracy and power efficiency. Similar to neural networks, one approach is to prune the HDC model using a mask, retaining the most crucial HDC dimensions postencoding [25]. Dimension importance can be determined by aggregating all class hypervectors into one and sorting the values across all dimensions. Notably, direct reduction of the encoding dimension should be avoided, as it can degrade HDC’s expressive capability and end up with corruption. Fig. 7 (left) visually demonstrates the impact of masking in supervised HDC tasks: retaining the top 6000 bits incurs only a 3% accuracy loss compared to using the full 10K-bit precision. While the concept of masking has been employed in prior HDC studies [25], they are not directly applicable to LifeHD due to their offline training setting with iid data. With streaming non-iid data in LifeHD, the set of observed cluster HVs may only represent a subset of the potential classes, and the less significant bits could become crucial as new classes are introduced. We introduce LifeHDa, which enhances LifeHD through an adaptive masking approach applied to all cluster HVs in working and long-term memory. Let 𝐷𝑎 represent the target dimension for reduction. The rationale behind LifeHDa is depicted in Fig. 7 (right). Whenever an original LifeHD detects novelty, we temporarily revert to the full dimension 𝐷 for 2 batches, which is sufficient for LifeHD to consolidate new patterns in its memory. After these two batches, we assess the long-term memory cluster HVs by aggregating them and ranking the dimensions, and derive a mask retaining 𝐷𝑎 dimensions with the largest absolute values. This mask is then applied to the following batches of 𝜙 (𝑋) immediately after encoding, up until the next novelty is detected. Novelty detection is executed with the masked hypervectors. Importantly, LifeHDa can utilize the same novelty detection sensitivity as LifeHD since the most significant dimensions dominate the similarity check. In other words, the similarity results in LifeHDa using 𝐷𝑎 dimension are similar as using the full dimension. LifeHDa offers an adaptive HDC model pruning interface with minimal accuracy loss and overhead. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Authors: Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction Abstract and 1. Introduction 2 Related Work 2 Related Work 3 Background on HDC 3 Background on HDC 4 Problem Definition 4 Problem Definition 5 LifeDH 5 LifeDH 6 Variants of LifeHD 6 Variants of LifeHD 7 Evaluation of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 10 Conclusion, Acknowledgments, and References 6 VARIANTS OF LIFEHD While LifeHD is designed to cater to general IoT applications with streaming input and without supervision, real-world scenarios may vary. Some scenarios might have a few labeled samples in addition to the unlabeled stream, while others may require operation within strict power constraints. LifeHD offers extensibility to address these diverse needs. In this section, we introduce two software-based extensions: LifeHDsemi, which adds a separate processing path to manage labeled samples, and LifeHDa, which adaptively prunes the HDC model using masking to handle low-power scenarios. 6.1 LifeHDsemi 6.2 LifeHDa While HDC computation is typically lightweight, there may be instances of energy scarcity (e.g., when powered by a solar panel) that call for a balance between accuracy and power efficiency. Similar to neural networks, one approach is to prune the HDC model using a mask, retaining the most crucial HDC dimensions postencoding [25]. Dimension importance can be determined by aggregating all class hypervectors into one and sorting the values across all dimensions. Notably, direct reduction of the encoding dimension should be avoided, as it can degrade HDC’s expressive capability and end up with corruption. Fig. 7 (left) visually demonstrates the impact of masking in supervised HDC tasks: retaining the top 6000 bits incurs only a 3% accuracy loss compared to using the full 10K-bit precision. While the concept of masking has been employed in prior HDC studies [25], they are not directly applicable to LifeHD due to their offline training setting with iid data. With streaming non-iid data in LifeHD, the set of observed cluster HVs may only represent a subset of the potential classes, and the less significant bits could become crucial as new classes are introduced. We introduce LifeHDa, which enhances LifeHD through an adaptive masking approach applied to all cluster HVs in working and long-term memory. Let 𝐷𝑎 represent the target dimension for reduction. The rationale behind LifeHDa is depicted in Fig. 7 (right). Whenever an original LifeHD detects novelty, we temporarily revert to the full dimension 𝐷 for 2 batches, which is sufficient for LifeHD to consolidate new patterns in its memory. After these two batches, we assess the long-term memory cluster HVs by aggregating them and ranking the dimensions, and derive a mask retaining 𝐷𝑎 dimensions with the largest absolute values. This mask is then applied to the following batches of 𝜙 (𝑋) immediately after encoding, up until the next novelty is detected. Novelty detection is executed with the masked hypervectors. Importantly, LifeHDa can utilize the same novelty detection sensitivity as LifeHD since the most significant dimensions dominate the similarity check. In other words, the similarity results in LifeHDa using 𝐷𝑎 dimension are similar as using the full dimension. LifeHDa offers an adaptive HDC model pruning interface with minimal accuracy loss and overhead. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. available on arxiv

Lifelong Intelligence Beyond the Edge using Hyperdimensional Computing: Variants of LifeHD

Authors:
(1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu);
(2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu);
(3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx);
(4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu);
(5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction 2 Related Work 3 Background on HDC 4 Problem Definition 5 LifeDH 6 Variants of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 5 LIFEHD In this section, we present the design of LifeHD, the first unsupervised HDC framework for lifelong learning in general edge IoT applications. Compared to operating in the original data space, HDC improves pattern separability through sparsity and high dimensionality, making it more resilient against catastrophic forgetting [52]. LifeHD preserves the advantages of HDC in computational efficiency and lifelong learning, while handling the input of unlabeled streaming data, which has not been achieved in previous work [18, 21, 22, 41, 65]. 5.1 LifeHD Overview Fig. 4 gives an overview of how LifeHD works. The first step is HDC encoding of data into hypervectors as described in Sec. 3. Training samples 𝑋 are organized into batches of size 𝑏𝑆𝑖𝑧𝑒 and input into an optional fixed NN for feature extraction (e.g. for images and sound) and the encoding module. The encoded hypervectors 𝜙 (𝑋) are input to LifeHD’s two-tier memory design inspired by cognitive science studies [5], consisting of working memory and long-term memory. This memory system intelligently and dynamically manages historical patterns, stored as hypervectors and referred to as cluster HVs. As shown in Fig. 4, the working memory is designed with three components: novelty detection, cluster HV update and cluster HV merge. 𝜙 (𝑋) is first input into novelty detection step (1). An insertion to the cluster HVs is made if a novelty flag is raised, otherwise 𝜙 (𝑋) updates the existing cluster HVs (2). The third component, cluster HV merge (3), retrieves the cluster HVs from long-term memory, and merges similar cluster HVs into a supercluster via a novel spectral clustering-based merging algorithm [59]. The interaction between working and long-term memory happens as commonly encountered cluster HVs are copied to long-term memory, which we call consolidation (4). Finally, when the size limit of either working or long-term memory is reached, the least recently used cluster HVs are forgotten (5). ecently used cluster HVs are forgotten (○5 ). All modules in LifeHD work collaboratively, making it adaptive and robust to continuously changing streams without relying on any form of prior knowledge. For example, in scenarios of distribution drift, LifeHD may generate new cluster HVs upon encountering drifted samples initially, which can later be merged into coarse clusters. This approach ensures that LifeHD can efficiently capture and retain historical patterns. In the following, we discuss more details about the major components of LifeHD: novelty detection (Sec. 5.2), cluster HV update (Sec. 5.3), and cluster HV merging (Sec. 5.4). We summarize the important notations used in this paper in Table 1. 5.2 Novelty Detection The initial novelty detection step (1 in Fig. 4) is crucial for identifying emerging patterns in the environment. SupposeM = {𝑚1, ...,𝑚𝑀} is the set of cluster HVs stored in the working memory. We gauge the "radius" of each cluster by tracking two scalars for each cluster HV 𝑖: 𝜇𝑖 and 𝜎ˆ𝑖 , which represent the mean cosine difference and standard difference between the cluster HV and its assigned inputs. Given 𝜙 (𝑋), we first identify the most similar cluster HV, denoted by 𝑗. LifeHD marks 𝜙 (𝑋) as “novel" if it substantially differs from its nearest cluster HV. Specifically, this dissimilarity is measured by comparing cos(𝜙 (𝑋),𝑚𝑗) with a threshold based on the historical distance distribution of cluster HV 𝑗: The hyperparameter 𝛾 fine-tunes the sensitivity to novelties. LifeHD recognizes new 𝜙 (𝑋) as prototypes and inserts them into the working memory. When reaching its size limit 𝑀, the working memory experiences forgetting (○5 in Fig. 4). The least recently used (LRU) cluster HV, represented by 𝐿𝑅𝑈 = argmin𝑀 𝑖=1 𝑝𝑖 , is replaced. Here 𝑝 corresponds to the latest batch index where the cluster HV was accessed. A similar forgetting mechanism is configured for the long-term memory, where the last batch accessed is marked with 𝑞. 5.3 Cluster HV Update If novelty is not detected, indicating that 𝜙 (𝑋) closely matches cluster HV 𝑗, we proceed to update the cluster HV and its associated information (2 in Fig. 4). This update process involves bundling 𝜙 (𝑋) with cluster HV 𝑚𝑗 , akin to how class hypervectors are updated as described in Sec. 3, and updated 𝜇𝑗 and 𝜎ˆ𝑗 with their moving average The hyperparameter 𝛼 adjusts the balance between historical and recent inputs, where a higher 𝛼 gives more weight to recent samples. Properly maintaining 𝜇𝑗 and 𝜎ˆ𝑗 is vital for tracking the “radius” of each cluster HV, affecting future novelty detection. We also increase the hit frequency ℎ𝑖𝑡𝑗 and refresh 𝑝𝑗 with current batch index 𝑖𝑑𝑥. ℎ𝑖𝑡𝑗 is further used to compared with a predetermined threshold ℎ𝑖𝑡𝑡ℎ to decide when a working memory cluster HV appears sufficiently frequently to be consolidated to long-term memory (4 in Fig. 4). 𝑝𝑗 determines forgetting as described in the previous section. With this lightweight approach, LifeHD continually records temporal cluster HVs from the environment, while the most prominent cluster HVs are transferred to long-term memory. 5.4 Cluster HV Merging Cluster HV merge (3 in Fig. 4) has the dual benefit of reducing memory use and of elucidating underlying similarity structure in the data. Intuitively, a group of cluster HVs can be merged if they are similar to each other and dissimilar from other cluster HVs. For instance, one might merge the cluster HVs for Bulldog and Chihuahua into a single “Dog” cluster HV, that remains distinct from the cluster HV for “Tabby Cat”. To merge the cluster HVs, we first construct a similarity graph defined over the cluster HVs from the long-term memory. The cluster HVs correspond to nodes, and a pair of cluster HVs are connected by an edge if they are sufficiently similar. We then merge the cluster HVs by computing a particular type of cut in the graph in a manner similar to spectral clustering [40]. This graph based formalism for clustering is able to capture complex types of cluster geometry and often substantially outperforms simpler approaches like K-Means [59]. We detail the steps of cluster HV merging in LifeHD below, while Fig. 5 offers an illustrative overview. Step 2: Decomposition. We compute the Laplacian𝑊 = 𝐷 −𝐴, where 𝐷 is the diagonal matrix in which 𝐷𝑖𝑖 = Í 𝑗 𝐴𝑖𝑗 . We then compute the eigenvalues 𝜆1, .., 𝜆𝐿, sorted in increasing order, and eigenvectors 𝜈1, ..., 𝜈𝐿 of 𝑊. Step 3: Grouping. We infer 𝑘 = max𝑖∈ [𝐿] 𝜆𝑖 ≤ 𝑔𝑢𝑏, and merge the cluster HVs by running K-Means on 𝜈1, ..., 𝜈𝑘 . The upperbound 𝑔𝑢𝑏 is a hyperparameter that adjusts the granularity of merging, with a smaller 𝑔𝑢𝑏 leading to smaller 𝑘 thus encouraging merging more aggressively. Our merging approach is formally grounded, as discussed in [59]. It is a well-known fact that the eigenvectors of 𝑊 encode information about the connected components of G. When G has 𝑘 connected components, the eigenvalues 𝜆1 = 𝜆2 = ... = 𝜆𝑘 = 0. To recover these components, K-Means clustering on 𝜈1, ..., 𝜈𝑘 can be employed, as explained in [59]. However, practical scenarios may have a few inter-component edges that should ideally be distinct. For instance, when the similarity threshold is imprecisely set, erroneous edges may appear in the graph, causing 𝜆1, ..., 𝜆𝑘 to be only approximately zero. Our merging approach is designed to handle this situation by introducing 𝑔𝑢𝑏. The cluster HV merging is evaluated every 𝑓𝑚𝑒𝑟𝑔𝑒 batches, where 𝑓𝑚𝑒𝑟𝑔𝑒 is a hyperparameter that controls the trade-off between merging performance and computational latency. Both 𝑔𝑢𝑏 and 𝑓𝑚𝑒𝑟𝑔𝑒 are analyzed in Sec. 7.8, along with other key hyperparameters in LifeHD. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Authors: Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction Abstract and 1. Introduction 2 Related Work 2 Related Work 3 Background on HDC 3 Background on HDC 4 Problem Definition 4 Problem Definition 5 LifeDH 5 LifeDH 6 Variants of LifeHD 6 Variants of LifeHD 7 Evaluation of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 10 Conclusion, Acknowledgments, and References 5 LIFEHD In this section, we present the design of LifeHD, the first unsupervised HDC framework for lifelong learning in general edge IoT applications. Compared to operating in the original data space, HDC improves pattern separability through sparsity and high dimensionality, making it more resilient against catastrophic forgetting [52]. LifeHD preserves the advantages of HDC in computational efficiency and lifelong learning, while handling the input of unlabeled streaming data, which has not been achieved in previous work [18, 21, 22, 41, 65]. 5.1 LifeHD Overview Fig. 4 gives an overview of how LifeHD works. The first step is HDC encoding of data into hypervectors as described in Sec. 3. Training samples 𝑋 are organized into batches of size 𝑏𝑆𝑖𝑧𝑒 and input into an optional fixed NN for feature extraction (e.g. for images and sound) and the encoding module. The encoded hypervectors 𝜙 (𝑋) are input to LifeHD’s two-tier memory design inspired by cognitive science studies [5], consisting of working memory and long-term memory. This memory system intelligently and dynamically manages historical patterns, stored as hypervectors and referred to as cluster HVs. As shown in Fig. 4, the working memory is designed with three components: novelty detection, cluster HV update and cluster HV merge. 𝜙 (𝑋) is first input into novelty detection step (1). An insertion to the cluster HVs is made if a novelty flag is raised, otherwise 𝜙 (𝑋) updates the existing cluster HVs (2). The third component, cluster HV merge (3), retrieves the cluster HVs from long-term memory, and merges similar cluster HVs into a supercluster via a novel spectral clustering-based merging algorithm [59]. The interaction between working and long-term memory happens as commonly encountered cluster HVs are copied to long-term memory, which we call consolidation (4). Finally, when the size limit of either working or long-term memory is reached, the least recently used cluster HVs are forgotten (5). ecently used cluster HVs are forgotten (○5 ). All modules in LifeHD work collaboratively, making it adaptive and robust to continuously changing streams without relying on any form of prior knowledge. For example, in scenarios of distribution drift, LifeHD may generate new cluster HVs upon encountering drifted samples initially, which can later be merged into coarse clusters. This approach ensures that LifeHD can efficiently capture and retain historical patterns. In the following, we discuss more details about the major components of LifeHD: novelty detection (Sec. 5.2), cluster HV update (Sec. 5.3), and cluster HV merging (Sec. 5.4). We summarize the important notations used in this paper in Table 1. 5.2 Novelty Detection The initial novelty detection step (1 in Fig. 4) is crucial for identifying emerging patterns in the environment. SupposeM = {𝑚1, ...,𝑚𝑀} is the set of cluster HVs stored in the working memory. We gauge the "radius" of each cluster by tracking two scalars for each cluster HV 𝑖: 𝜇𝑖 and 𝜎ˆ𝑖 , which represent the mean cosine difference and standard difference between the cluster HV and its assigned inputs. Given 𝜙 (𝑋), we first identify the most similar cluster HV, denoted by 𝑗. LifeHD marks 𝜙 (𝑋) as “novel" if it substantially differs from its nearest cluster HV. Specifically, this dissimilarity is measured by comparing cos(𝜙 (𝑋),𝑚𝑗) with a threshold based on the historical distance distribution of cluster HV 𝑗: The hyperparameter 𝛾 fine-tunes the sensitivity to novelties. LifeHD recognizes new 𝜙 (𝑋) as prototypes and inserts them into the working memory. When reaching its size limit 𝑀, the working memory experiences forgetting (○5 in Fig. 4). The least recently used (LRU) cluster HV, represented by 𝐿𝑅𝑈 = argmin𝑀 𝑖=1 𝑝𝑖 , is replaced. Here 𝑝 corresponds to the latest batch index where the cluster HV was accessed. A similar forgetting mechanism is configured for the long-term memory, where the last batch accessed is marked with 𝑞. 5.3 Cluster HV Update If novelty is not detected, indicating that 𝜙 (𝑋) closely matches cluster HV 𝑗, we proceed to update the cluster HV and its associated information (2 in Fig. 4). This update process involves bundling 𝜙 (𝑋) with cluster HV 𝑚𝑗 , akin to how class hypervectors are updated as described in Sec. 3, and updated 𝜇𝑗 and 𝜎ˆ𝑗 with their moving average The hyperparameter 𝛼 adjusts the balance between historical and recent inputs, where a higher 𝛼 gives more weight to recent samples. Properly maintaining 𝜇𝑗 and 𝜎ˆ𝑗 is vital for tracking the “radius” of each cluster HV, affecting future novelty detection. We also increase the hit frequency ℎ𝑖𝑡𝑗 and refresh 𝑝𝑗 with current batch index 𝑖𝑑𝑥. ℎ𝑖𝑡𝑗 is further used to compared with a predetermined threshold ℎ𝑖𝑡𝑡ℎ to decide when a working memory cluster HV appears sufficiently frequently to be consolidated to long-term memory (4 in Fig. 4). 𝑝𝑗 determines forgetting as described in the previous section. With this lightweight approach, LifeHD continually records temporal cluster HVs from the environment, while the most prominent cluster HVs are transferred to long-term memory. 5.4 Cluster HV Merging Cluster HV merge (3 in Fig. 4) has the dual benefit of reducing memory use and of elucidating underlying similarity structure in the data. Intuitively, a group of cluster HVs can be merged if they are similar to each other and dissimilar from other cluster HVs. For instance, one might merge the cluster HVs for Bulldog and Chihuahua into a single “Dog” cluster HV, that remains distinct from the cluster HV for “Tabby Cat”. To merge the cluster HVs, we first construct a similarity graph defined over the cluster HVs from the long-term memory. The cluster HVs correspond to nodes, and a pair of cluster HVs are connected by an edge if they are sufficiently similar. We then merge the cluster HVs by computing a particular type of cut in the graph in a manner similar to spectral clustering [40]. This graph based formalism for clustering is able to capture complex types of cluster geometry and often substantially outperforms simpler approaches like K-Means [59]. We detail the steps of cluster HV merging in LifeHD below, while Fig. 5 offers an illustrative overview. Step 2: Decomposition. We compute the Laplacian𝑊 = 𝐷 −𝐴, where 𝐷 is the diagonal matrix in which 𝐷𝑖𝑖 = Í 𝑗 𝐴𝑖𝑗 . We then compute the eigenvalues 𝜆1, .., 𝜆𝐿, sorted in increasing order, and eigenvectors 𝜈1, ..., 𝜈𝐿 of 𝑊. Step 2: Decomposition. Step 3: Grouping. We infer 𝑘 = max𝑖∈ [𝐿] 𝜆𝑖 ≤ 𝑔𝑢𝑏, and merge the cluster HVs by running K-Means on 𝜈1, ..., 𝜈𝑘 . The upperbound 𝑔𝑢𝑏 is a hyperparameter that adjusts the granularity of merging, with a smaller 𝑔𝑢𝑏 leading to smaller 𝑘 thus encouraging merging more aggressively. Our merging approach is formally grounded, as discussed in [59]. It is a well-known fact that the eigenvectors of 𝑊 encode information about the connected components of G. When G has 𝑘 connected components, the eigenvalues 𝜆1 = 𝜆2 = ... = 𝜆𝑘 = 0. To recover these components, K-Means clustering on 𝜈1, ..., 𝜈𝑘 can be employed, as explained in [59]. However, practical scenarios may have a few inter-component edges that should ideally be distinct. For instance, when the similarity threshold is imprecisely set, erroneous edges may appear in the graph, causing 𝜆1, ..., 𝜆𝑘 to be only approximately zero. Our merging approach is designed to handle this situation by introducing 𝑔𝑢𝑏. The cluster HV merging is evaluated every 𝑓𝑚𝑒𝑟𝑔𝑒 batches, where 𝑓𝑚𝑒𝑟𝑔𝑒 is a hyperparameter that controls the trade-off between merging performance and computational latency. Both 𝑔𝑢𝑏 and 𝑓𝑚𝑒𝑟𝑔𝑒 are analyzed in Sec. 7.8, along with other key hyperparameters in LifeHD. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. available on arxiv

Lifelong Intelligence Beyond the Edge using Hyperdimensional Computing: LifeDH

Authors:
(1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu);
(2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu);
(3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx);
(4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu);
(5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction 2 Related Work 3 Background on HDC 4 Problem Definition 5 LifeDH 6 Variants of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 4 PROBLEM DEFINITION Before diving into our method, we first rigorously formulate the unsupervised lifelong learning problem using streaming sources, driven by real-world IoT applications. Streaming Data. To represent continuously changing environment, we assume a well-known class-incremental model in lifelong learning, in which new classes emerge in a sequential manner [46]. We also allow data distribution shift within one class. This setting models a scenario in which a device is continuously sampling data while the surrounding environment may change implicitly over time, e.g., the self-driving vehicle as shown in Fig. 1. We require that all samples appear only once (i.e., single-pass streams). Learning Protocol. Our goal is to build a classification algorithm that maps X → Y. For evaluation, we use the common evaluation protocol in state-of-the-art lifelong learning works [13, 14, 54], in which we construct an iid dataset E = {(𝑋𝑘 , 𝑦𝑘 )} for periodic testing, by sampling labeled examples from each class in a manner that preserves the overall (im)balance between the classes. Note, that even when one class has not appeared in the training data stream, it is always included in E. Hence E is a global view of all classes that can potentially exist in the environment. Unsupervised Clustering Accuracy. Since we do not give class labels or the total number of classes during training, the predicted label can be different from the ground-truth label. Therefore, for evaluation metric, we cannot adopt the simple prediction accuracy that requires exact label matching. Instead, we employ a widely used clustering metric known as unsupervised clustering accuracy (ACC) [63], which mirrors the conventional accuracy evaluation but within an unsupervised context. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Authors: Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction Abstract and 1. Introduction 2 Related Work 2 Related Work 3 Background on HDC 3 Background on HDC 4 Problem Definition 4 Problem Definition 5 LifeDH 5 LifeDH 6 Variants of LifeHD 6 Variants of LifeHD 7 Evaluation of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 10 Conclusion, Acknowledgments, and References 4 PROBLEM DEFINITION Before diving into our method, we first rigorously formulate the unsupervised lifelong learning problem using streaming sources, driven by real-world IoT applications. Streaming Data. To represent continuously changing environment, we assume a well-known class-incremental model in lifelong learning, in which new classes emerge in a sequential manner [46]. We also allow data distribution shift within one class. This setting models a scenario in which a device is continuously sampling data while the surrounding environment may change implicitly over time, e.g., the self-driving vehicle as shown in Fig. 1. We require that all samples appear only once (i.e., single-pass streams). Streaming Data. Learning Protocol. Our goal is to build a classification algorithm that maps X → Y. For evaluation, we use the common evaluation protocol in state-of-the-art lifelong learning works [13, 14, 54], in which we construct an iid dataset E = {(𝑋𝑘 , 𝑦𝑘 )} for periodic testing, by sampling labeled examples from each class in a manner that preserves the overall (im)balance between the classes. Note, that even when one class has not appeared in the training data stream, it is always included in E. Hence E is a global view of all classes that can potentially exist in the environment. Learning Protocol. Unsupervised Clustering Accuracy. Since we do not give class labels or the total number of classes during training, the predicted label can be different from the ground-truth label. Therefore, for evaluation metric, we cannot adopt the simple prediction accuracy that requires exact label matching. Instead, we employ a widely used clustering metric known as unsupervised clustering accuracy (ACC) [63], which mirrors the conventional accuracy evaluation but within an unsupervised context. Unsupervised Clustering Accuracy. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. available on arxiv

Lifelong Intelligence Beyond the Edge using Hyperdimensional Computing: Problem Definition

Authors:
(1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu);
(2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu);
(3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx);
(4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu);
(5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction 2 Related Work 3 Background on HDC 4 Problem Definition 5 LifeDH 6 Variants of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 3 BACKGROUND ON HDC Hyperdimensional Computing (HDC) is an emerging paradigm for information processing from the cognitive-neuroscience literature [24]. In HDC, all computation is performed on low-precision and distributed representations of data that accord naturally with highly parallel and low-energy hardware. The encoding function 𝜙 : X → H embeds data from its ambient representation into HD-space. In general, encoding should preserve some meaningful notion of similarity between input points in the sense that 𝜙 (𝑥) · 𝜙 (𝑥 ′ ) ≈ 𝑘 (𝑥, 𝑥′ ), where 𝑘 is some similarity function of interest on X. In this paper, we use spatiotemporal encoding for time series sensor data, and HDnn for more complex data, such as images, which we explain in the following. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Authors: Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction Abstract and 1. Introduction 2 Related Work 2 Related Work 3 Background on HDC 3 Background on HDC 4 Problem Definition 4 Problem Definition 5 LifeDH 5 LifeDH 6 Variants of LifeHD 6 Variants of LifeHD 7 Evaluation of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 10 Conclusion, Acknowledgments, and References 3 BACKGROUND ON HDC Hyperdimensional Computing (HDC) is an emerging paradigm for information processing from the cognitive-neuroscience literature [24]. In HDC, all computation is performed on low-precision and distributed representations of data that accord naturally with highly parallel and low-energy hardware. The encoding function 𝜙 : X → H embeds data from its ambient representation into HD-space. In general, encoding should preserve some meaningful notion of similarity between input points in the sense that 𝜙 (𝑥) · 𝜙 (𝑥 ′ ) ≈ 𝑘 (𝑥, 𝑥′ ), where 𝑘 is some similarity function of interest on X. In this paper, we use spatiotemporal encoding for time series sensor data, and HDnn for more complex data, such as images, which we explain in the following. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. available on arxiv

Lifelong Intelligence Beyond the Edge using Hyperdimensional Computing: Background on HDC

Authors:
(1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu);
(2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu);
(3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx);
(4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu);
(5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction 2 Related Work 3 Background on HDC 4 Problem Definition 5 LifeDH 6 Variants of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 2 RELATED WORK Lifelong and On-Device Learning. Lifelong learning (or continual learning) is a large and active area of research in the broader machine learning community. Catastrophic forgetting is a major challenge in lifelong learning, and refers to a commonly observed empirical phenomenon in which updating certain machine learning models with new data severely degrades their ability to perform previously learned tasks [36]. Previous works proposed techniques such as dynamic architecture [31, 49], regularization by penalizing important weights [28, 66], knowledge distillation from past models [14] and experience replay using a memory buffer [35, 58]. The lifelong learning literature has examined a wide range of problem settings, ranging from the fully supervised case, in which tasks and class labels are provided, and the fully unsupervised case without any labels and prior knowledge [13, 57]. However, all of these works are based on deep NNs and require backpropagation, which is problematic for resource-constrained devices. Neurally-inspired lightweight algorithms have recently been proposed for lifelong learning applications. FlyModel [52] and SDMLP [8] use sparse coding and associative memory for lifelong learning. However, both approaches assume full supervision. STAM [54] is an expandable memory architecture for unsupervised lifelong learning, using layered receptive fields and a two-tier memory hierarchy. It learns via online centroid-based clustering pipeline, novelty detection and memory updates. Nevertheless, the memory in STAM is solely dedicated to image storage, while our LifeHD additionally emphasizes merging past patterns into coarse groups and shows more effective learning performance. Recent works optimize the resource usage of on-device training via pruning and quantization [34, 45], tuning partial weights [9, 47], memory profiling and optimization [15, 61, 64], as well as growing the NN on the fly [67]. All these works optimize training given resource constraints and do not focus on lifelong learning. They are orthogonal to the contribution of LifeHD which focuses on adaptive and continual training. LifeHD can be further optimized by combining with such techniques. Hyperdimensional Computing. HDC has garnered substantial interest from the computer hardware community as an energyefficient and low-latency approach to learning, and has been successfully applied to problems such as human activity recognition [27], voice recognition [23], image recognition [11, 65], to name a few. The large majority of literature on HDC has focused on using the technique to perform supervised classification tasks. Among the limited literature for weakly-supervised learning with HDC, HDCluster [21] enabled unsupervised clustering in HDC with a new algorithm that is similar to K-Means. SemiHD [22] is a semi-supervised learning framework using HDC with iterative self-labeling. Hyperseed [41], C-FSCIL [18] and FSL-HD [65] adopted HDC or similar vector symbolic architectures (VSA) for unsupervised or few-shot learning. All above works did not consider the lifelong aspect and used offline training on a static dataset. To the best of the authors’ knowledge, LifeHD is the first work that designs and deploys lifelong learning in edge IoT applications especially with zero or minimal amount of labels. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Authors: Authors: (1) Xiaofan Yu, University of California San Diego, La Jolla, California, USA (x1yu@ucsd.edu); (2) Anthony Thomas, University of California San Diego, La Jolla, California, USA (ahthomas@ucsd.edu); (3) Ivannia Gomez Moreno, CETYS University, Campus Tijuana, Tijuana, Mexico (ivannia.gomez@cetys.edu.mx); (4) Louis Gutierrez, University of California San Diego, La Jolla, California, USA (l8gutierrez@ucsd.edu); (5) Tajana Šimunić Rosing, University of California San Diego, La Jolla, USA (tajana@ucsd.edu). Table of Links Abstract and 1. Introduction Abstract and 1. Introduction 2 Related Work 2 Related Work 3 Background on HDC 3 Background on HDC 4 Problem Definition 4 Problem Definition 5 LifeDH 5 LifeDH 6 Variants of LifeHD 6 Variants of LifeHD 7 Evaluation of LifeHD 7 Evaluation of LifeHD 8 Evaluation of LifeHD semi and LifeHDa 8 Evaluation of LifeHD semi and LifeHDa 9 Discussions and Future Works 9 Discussions and Future Works 10 Conclusion, Acknowledgments, and References 10 Conclusion, Acknowledgments, and References 2 RELATED WORK 2 RELATED WORK Lifelong and On-Device Learning. Lifelong learning (or continual learning) is a large and active area of research in the broader machine learning community. Catastrophic forgetting is a major challenge in lifelong learning, and refers to a commonly observed empirical phenomenon in which updating certain machine learning models with new data severely degrades their ability to perform previously learned tasks [36]. Previous works proposed techniques such as dynamic architecture [31, 49], regularization by penalizing important weights [28, 66], knowledge distillation from past models [14] and experience replay using a memory buffer [35, 58]. The lifelong learning literature has examined a wide range of problem settings, ranging from the fully supervised case, in which tasks and class labels are provided, and the fully unsupervised case without any labels and prior knowledge [13, 57]. However, all of these works are based on deep NNs and require backpropagation, which is problematic for resource-constrained devices. Lifelong and On-Device Learning. Neurally-inspired lightweight algorithms have recently been proposed for lifelong learning applications. FlyModel [52] and SDMLP [8] use sparse coding and associative memory for lifelong learning. However, both approaches assume full supervision. STAM [54] is an expandable memory architecture for unsupervised lifelong learning, using layered receptive fields and a two-tier memory hierarchy. It learns via online centroid-based clustering pipeline, novelty detection and memory updates. Nevertheless, the memory in STAM is solely dedicated to image storage, while our LifeHD additionally emphasizes merging past patterns into coarse groups and shows more effective learning performance. Recent works optimize the resource usage of on-device training via pruning and quantization [34, 45], tuning partial weights [9, 47], memory profiling and optimization [15, 61, 64], as well as growing the NN on the fly [67]. All these works optimize training given resource constraints and do not focus on lifelong learning. They are orthogonal to the contribution of LifeHD which focuses on adaptive and continual training. LifeHD can be further optimized by combining with such techniques. Hyperdimensional Computing. HDC has garnered substantial interest from the computer hardware community as an energyefficient and low-latency approach to learning, and has been successfully applied to problems such as human activity recognition [27], voice recognition [23], image recognition [11, 65], to name a few. The large majority of literature on HDC has focused on using the technique to perform supervised classification tasks. Among the limited literature for weakly-supervised learning with HDC, HDCluster [21] enabled unsupervised clustering in HDC with a new algorithm that is similar to K-Means. SemiHD [22] is a semi-supervised learning framework using HDC with iterative self-labeling. Hyperseed [41], C-FSCIL [18] and FSL-HD [65] adopted HDC or similar vector symbolic architectures (VSA) for unsupervised or few-shot learning. All above works did not consider the lifelong aspect and used offline training on a static dataset. To the best of the authors’ knowledge, LifeHD is the first work that designs and deploys lifelong learning in edge IoT applications especially with zero or minimal amount of labels. Hyperdimensional Computing. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. available on arxiv

Lifelong Intelligence Beyond the Edge using Hyperdimensional Computing: Related Work

Lifelong Intelligence Beyond the Edge using Hyperdimensional Computing: Abstract and Introduction

Visualizing Colors: The Gradient Oscillation Hyperspace

ভিজ্যুয়ালাইজিং কালার: গ্রেডিয়েন্ট অসিলেশন হাইপারস্পেস

Farben visualisieren: Der Gradientenoszillations-Hyperraum

Visualización de colores: el hiperespacio de oscilación de gradiente

Visualisation des couleurs : l'hyperespace d'oscillation de dégradé

रंगों का दृश्यांकन: ग्रेडिएंट ऑसिलेशन हाइपरस्पेस

色の視覚化: 勾配振動ハイパースペース

색상 시각화: 그라데이션 진동 초공간

Visualizando Cores: O Hiperespaço de Oscilação de Gradiente

Визуализация цветов: гиперпространство градиентных колебаний

Renkleri Görselleştirmek: Degrade Salınım Hiperuzayı

Trực quan hóa màu sắc: Siêu không gian dao động chuyển màu

可视化颜色：梯度振荡超空间

Businesses and users are going to use technology only if they can trust it,”Satya Nadella” If water is life, what about data in tech? It’s the oil, the diamond, the precious asset we all need. With data, everything we do in tech is meaningful! Without data, there will be no black hat hackers seeking information to steal online and offline! With an abundance of generative AI tools, the hype is still present, some people claim writers, graphic designers, and developers, will be replaced in the coming years. Every day new tools come into the market and the hype continues, what’s next? The generative AI dilemma, hype vs trust What if the hype stops tomorrow? As soon as people use generative AI tools, trust will become a great concern! The way some tools have been marketed in the industry is not what people expect to achieve their tasks in their daily routines. Everything is hackable! some studies have shown how chatGPT could be tricked by leaking data and a paper explains how an adversary could extract gigabytes of data from open source models and closed models like ChatGPT. The worry is no matter the compliance we have it’s not guaranteed users will be protected at 100 %, some of the well-known in the industry are HIPAA, PCI-DSS, GDPR … compliance is a great way to prove that companies care about privacy and respect a set of rules to protect users from data theft, unfortunately, it’s not guaranteed companies will be unhackable. in the generative AI era, companies should ensure users their data is secure and nobody will steal them easily, this is the way to go beyond the hype. Cyberattacks can happen every day, the most protected companies will retain customers those with fewer means will lose astronomically, we are at the dawn of massive cyber attacks on generative AI tools, as soon as they can be tricked and with unaware users with less cybersecurity literacy the damage will cost millions! What’s next after the generative AI  hype? We need more trust, and users by categories should be educated weekly and monthly on the potential risks they face daily with generative AI tools. Developers need GDPR-driven services or apis to integrate into their for better security and they should always consider the secure-by-design approach at every step of the Software development lifecycle before shipping their apps, beyond the hype data is the precious asset we all need and we use it to improve lives or create damage! Privacy should not remain a slogan, by trusting a service we put our lives in the hands of a company in return we need better security by ensuring that our data no matter the day is safe nobody can steal and we will always remain secure. Imagine losing access to your favorite services online after your sensitive data was stolen online by hackers! For users who are not technical, they need continuous training, hackers change their techniques daily to perform large attacks and we never know who will be a victim, by changing the behavior of how they use generative AI tools they will improve their security on their side too, companies do their job, they are not responsible for all users behaviors! Recommendations for trust in the generative AI era Companies should invest in continuous training for developers from a technical and nontechnical side
Developers advocates should educate users with less technical skills on how to secure themselves using generative AI tools for better data privacy.
Compliance is not enough for data privacy, the secure by design approach should be the way to go before shipping any product.
We need a fruitful collaboration between several stakeholders for better data privacy and security. Conclusion The Hype will come to an end ,it’s time for more trust,users can’t be believe in the longrun in services they are afraid to use,privacy is precious and data is the fuel that moves the industry. Trust will be the new fuel,users observe the industry,some lose their data,some escape potential risks of sensitive data theft everyday and they don’t know which services is better than others amid the AI noise. A new era will come,at the end all we need is data privacy. Businesses and users are going to use technology only if they can trust it,”Satya Nadella” Businesses and users are going to use technology only if they can trust it,” Satya Nadella ” Satya Nadella If water is life, what about data in tech? It’s the oil, the diamond, the precious asset we all need. With data, everything we do in tech is meaningful! Without data, there will be no black hat hackers seeking information to steal online and offline! With an abundance of generative AI tools, the hype is still present, some people claim writers, graphic designers, and developers, will be replaced in the coming years. Every day new tools come into the market and the hype continues, what’s next? The generative AI dilemma, hype vs trust What if the hype stops tomorrow? As soon as people use generative AI tools, trust will become a great concern! The way some tools have been marketed in the industry is not what people expect to achieve their tasks in their daily routines. Everything is hackable! some studies have shown how chatGPT could be tricked by leaking data and a paper explains how an adversary could extract gigabytes of data from open source models and closed models like ChatGPT. chatGPT paper paper The worry is no matter the compliance we have it’s not guaranteed users will be protected at 100 %, some of the well-known in the industry are HIPAA , PCI-DSS , GDPR … HIPAA PCI-DSS GDPR compliance is a great way to prove that companies care about privacy and respect a set of rules to protect users from data theft, unfortunately, it’s not guaranteed companies will be unhackable. in the generative AI era, companies should ensure users their data is secure and nobody will steal them easily, this is the way to go beyond the hype. Cyberattacks can happen every day, the most protected companies will retain customers those with fewer means will lose astronomically, we are at the dawn of massive cyber attacks on generative AI tools, as soon as they can be tricked and with unaware users with less cybersecurity literacy the damage will cost millions! What’s next after the generative AI  hype? What’s next after the generative AI  hype? We need more trust, and users by categories should be educated weekly and monthly on the potential risks they face daily with generative AI tools. Developers need GDPR-driven services or apis to integrate into their for better security and they should always consider the secure-by-design approach at every step of the Software development lifecycle before shipping their apps, beyond the hype data is the precious asset we all need and we use it to improve lives or create damage! GDPR-driven services GDPR-driven services Privacy should not remain a slogan, by trusting a service we put our lives in the hands of a company in return we need better security by ensuring that our data no matter the day is safe nobody can steal and we will always remain secure. Imagine losing access to your favorite services online after your sensitive data was stolen online by hackers! For users who are not technical, they need continuous training, hackers change their techniques daily to perform large attacks and we never know who will be a victim, by changing the behavior of how they use generative AI tools they will improve their security on their side too, companies do their job, they are not responsible for all users behaviors! Recommendations for trust in the generative AI era Recommendations for trust in the generative AI era Companies should invest in continuous training for developers from a technical and nontechnical side Developers advocates should educate users with less technical skills on how to secure themselves using generative AI tools for better data privacy. Compliance is not enough for data privacy, the secure by design approach should be the way to go before shipping any product. We need a fruitful collaboration between several stakeholders for better data privacy and security. Companies should invest in continuous training for developers from a technical and nontechnical side Developers advocates should educate users with less technical skills on how to secure themselves using generative AI tools for better data privacy. Compliance is not enough for data privacy, the secure by design approach should be the way to go before shipping any product. We need a fruitful collaboration between several stakeholders for better data privacy and security. Conclusion The Hype will come to an end ,it’s time for more trust,users can’t be believe in the longrun in services they are afraid to use,privacy is precious and data is the fuel that moves the industry. Trust will be the new fuel,users observe the industry,some lose their data,some escape potential risks of sensitive data theft everyday and they don’t know which services is better than others amid the AI noise. A new era will come,at the end all we need is data privacy.

Generative AI : Hype, Hype, Hype, What's Next? 

Hypernative is set to implement its real-time threat detection and response solution across Stacks, Bitcoin's leading Layer 2 (L2) platform. This collaboration aims to strengthen the security infrastructure for applications and smart contracts built on Stacks, reflecting the growing need for advanced protection in complex blockchain environments. The integration of Hypernative's security solutions into Stacks will provide projects operating on this Bitcoin L2 with access to state-of-the-art cybersecurity tools. This includes: Real-time monitoring and threat detection


Assistance in configuring security monitoring and prevention workflows


Incident response support, including investigation and root cause analysis


Access to Hypernative's network for potential fund recovery in case of theft Gal Sagie, CEO and co-founder of Hypernative, emphasized the importance of this integration: "Stacks is on a mission to scale Bitcoin and make more advanced financial constructs such as decentralized finance possible on the network. With more code and more smart contracts comes a critical need to make sure they are safe. Hypernative is proud to secure the entire Stacks ecosystem in real time and help activate the Bitcoin economy." Hypernative's platform is designed to monitor both on-chain and off-chain data sources in real-time, capable of identifying over 200 risk types. The system employs advanced technologies including: Machine learning algorithms


Heuristics


Simulations


Graph-based analytics This comprehensive approach allows for rapid identification and response to potential security threats across various components of the Stacks ecosystem, including smart contracts, cross-chain bridges, and treasury management systems. The collaboration between Hypernative and Stacks comes at a crucial time for Bitcoin-based L2 solutions. As more sophisticated financial structures are built on top of Bitcoin, the demand for robust security measures increases proportionally. Dr. Elena Rodriguez, a blockchain security researcher at CryptoTech Institute, noted: "As we see more advanced financial constructs being built on top of Bitcoin, the security requirements naturally increase. It's a necessary evolution in the space." The implementation of these security measures aligns with Stacks' preparation for its "Nakamoto release," scheduled for later this year. This update promises to bring increased transaction speeds and improved finality to the network. Looking Ahead As the Bitcoin ecosystem continues to evolve beyond its original design as a peer-to-peer electronic cash system, the balance between innovation and security remains a key challenge. The collaboration between Hypernative and Stacks represents a proactive approach to addressing this challenge, potentially setting a precedent for other L2 solutions. The effectiveness of these measures will be proven over time, and the broader crypto community will be watching closely. The outcomes could significantly influence security strategies across the rapidly changing landscape of blockchain technology. Don’t forget to like and share the story! Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR. Hypernative is set to implement its real-time threat detection and response solution across Stacks , Bitcoin's leading Layer 2 (L2) platform. This collaboration aims to strengthen the security infrastructure for applications and smart contracts built on Stacks, reflecting the growing need for advanced protection in complex blockchain environments. Stacks The integration of Hypernative's security solutions into Stacks will provide projects operating on this Bitcoin L2 with access to state-of-the-art cybersecurity tools. This includes: Real-time monitoring and threat detection Assistance in configuring security monitoring and prevention workflows Incident response support, including investigation and root cause analysis Access to Hypernative's network for potential fund recovery in case of theft Real-time monitoring and threat detection Real-time monitoring and threat detection Assistance in configuring security monitoring and prevention workflows Assistance in configuring security monitoring and prevention workflows Incident response support, including investigation and root cause analysis Incident response support, including investigation and root cause analysis Access to Hypernative's network for potential fund recovery in case of theft Access to Hypernative's network for potential fund recovery in case of theft Gal Sagie, CEO and co-founder of Hypernative, emphasized the importance of this integration: "Stacks is on a mission to scale Bitcoin and make more advanced financial constructs such as decentralized finance possible on the network. With more code and more smart contracts comes a critical need to make sure they are safe. Hypernative is proud to secure the entire Stacks ecosystem in real time and help activate the Bitcoin economy." Hypernative's platform is designed to monitor both on-chain and off-chain data sources in real-time, capable of identifying over 200 risk types. The system employs advanced technologies including: Machine learning algorithms Heuristics Simulations Graph-based analytics Machine learning algorithms Machine learning algorithms Heuristics Heuristics Simulations Simulations Graph-based analytics Graph-based analytics This comprehensive approach allows for rapid identification and response to potential security threats across various components of the Stacks ecosystem, including smart contracts, cross-chain bridges, and treasury management systems. The collaboration between Hypernative and Stacks comes at a crucial time for Bitcoin-based L2 solutions. As more sophisticated financial structures are built on top of Bitcoin, the demand for robust security measures increases proportionally. Dr. Elena Rodriguez, a blockchain security researcher at CryptoTech Institute, noted: "As we see more advanced financial constructs being built on top of Bitcoin, the security requirements naturally increase. It's a necessary evolution in the space." The implementation of these security measures aligns with Stacks' preparation for its "Nakamoto release," scheduled for later this year. This update promises to bring increased transaction speeds and improved finality to the network. Looking Ahead As the Bitcoin ecosystem continues to evolve beyond its original design as a peer-to-peer electronic cash system, the balance between innovation and security remains a key challenge. The collaboration between Hypernative and Stacks represents a proactive approach to addressing this challenge, potentially setting a precedent for other L2 solutions. The effectiveness of these measures will be proven over time, and the broader crypto community will be watching closely. The outcomes could significantly influence security strategies across the rapidly changing landscape of blockchain technology. Don’t forget to like and share the story! Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR. Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program . HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR. Vested Interest Disclosure: business blogging program business blogging program

Hypernative Bolsters Bitcoin L2 Security as Stacks Ecosystem Gets Real-Time Protection

হাইপারনেটিভ বলস্টার বিটকয়েন L2 সিকিউরিটি হিসাবে স্ট্যাক ইকোসিস্টেম রিয়েল-টাইম সুরক্ষা পায়

Hypernative stärkt Bitcoin L2-Sicherheit, während das Stacks-Ökosystem Echtzeitschutz erhält

Hypernative refuerza la seguridad de Bitcoin L2 a medida que el ecosistema Stacks obtiene protección en tiempo real

Hypernative renforce la sécurité Bitcoin L2 alors que l'écosystème Stacks bénéficie d'une protection en temps réel

हाइपरनेटिव ने बिटकॉइन L2 सुरक्षा को मजबूत किया क्योंकि स्टैक इकोसिस्टम को वास्तविक समय की सुरक्षा मिली

Hypernative が Stacks エコシステムにリアルタイム保護を導入し、ビットコイン L2 セキュリティを強化

Hypernative는 스택 생태계가 실시간 보호를 받음에 따라 비트코인 L2 보안을 강화합니다.

Hypernative reforça a segurança L2 do Bitcoin à medida que o ecossistema Stacks obtém proteção em tempo real

Hypernative повышает безопасность биткойнов L2, поскольку экосистема стеков получает защиту в реальном времени

Hypernative, Stacks Ekosistemi Gerçek Zamanlı Koruma Aldıkça Bitcoin L2 Güvenliğini Güçlendiriyor

Hypernative tăng cường bảo mật Bitcoin L2 khi hệ sinh thái ngăn xếp được bảo vệ theo thời gian thực

Hypernative 增强比特币 L2 安全性，为 Stacks 生态系统提供实时保护

The Bitcoin Hypocrisy Repository - Keeping Track Of Mainstream Lies & Fake Narratives

I remember the day one of my good friends reached out and asked if I had heard about Nuklai. I simply told him, “Not really. That sounds like a topic I would study in my science class.” Fast forward six months, and here I am, building a Layer 1 blockchain network for Nuklai. Who would’ve thought? It is amazing where life takes you in a matter of months. I will get to the core of this blog about my journey into the world of HyperVMs, but before that, it is essential to know how I even got to this point. The Challenge of High Throughput Soon after I heard about Nuklai, I started to research the project to see what all the hype was about. I was really blown away by Nuklai’s vision to empower every user by allowing them to sell their data to other entities, shifting the power from big businesses to the users themselves. It is a simple yet profound vision because, to achieve it, much had to be done not just at the application level but also at the infrastructure and blockchain protocol levels. When dealing with massive amounts of data daily, the blockchain must handle an immense load that no current platform has achieved yet. People talk about Transactions Per Second (TPS), but few know what it truly takes to achieve such high TPS at the blockchain level. Yes, there are blockchains out there with high throughput, but often, they are very hacky and not true Layer 1s. Usually, the approach is to scale horizontally by creating Layer 2s, Layer 3s, sidechains, and whatnot. These solutions feel like patchwork to me because if you introduce an additional layer to your blockchain, increasing the TPS seems more like an afterthought than something built from the ground up. This is where HyperVMs come into play. HyperVMs, Unlocking High-Performance Blockchains But, what is a HyperVM? It’s essentially a virtual machine (VM) built using the HyperSDK framework. To understand HyperSDK, we first need to understand Avalanche. According to their documentation, Avalanche is an open-source platform for building decentralized applications in one interoperable, decentralized, and highly scalable ecosystem. Powered by a uniquely powerful consensus mechanism, Avalanche is the first ecosystem designed to accommodate the scale of global finance, with near-instant transaction finality. The Avalanche consensus mechanism is ridiculously fast, with less than 2 seconds of finality, meaning your transaction is processed and verified almost instantly. Avalanche also allows you to create your own Layer 1 blockchain (or subnet), and the best part is you are not confined to an Ethereum Virtual Machine (EVM) powered subnet. Despite its large developer community, the EVM is not lightweight and comes with a lot of baggage. On Avalanche, you can also launch your own subnet by developing your own VM in any language as long as it conforms to Avalanche’s Proto interfaces. This is where HyperSDK comes into play. Theoretically, you could build your own VM in any language you want, as Avalanche provides that freedom. However, it is also true that it is rather difficult and time-intensive to do from scratch. Even if you use Subnet EVM to get started, it is complex to keep it up-to-date as the core code changes frequently. You need to keep up with maintaining it, which can become resource-intensive. HyperSDK addresses this issue by making it faster, safer, and easier to launch your own optimized blockchain on an Avalanche subnet. It hides much of the complexity, allowing you to focus on making your project unique rather than worrying about transaction serialization efficiency or increasing transaction throughput. With HyperSDK, you can build your own blockchain runtime with 500-1000 lines of code and still process thousands of transactions per second without needing a large team of engineers. HelixVM: The 100,000 tx/s Powerhouse? At Nuklai, we built our own blockchain runtime, called ‘HelixVM,’ using the HyperSDK framework with around 200 lines of code in under a week. We added exciting features such as natively issuing tokens, bridging tokens from one subnet to another using Avalanche Warp Messaging (AWM), and developing a staking engine that lets validators register for staking and users delegate their NAI tokens to a validator of their choice to start claiming rewards. With some more additions and custom touches, we added everything with around 500 more lines of code in a couple of months. The journey to get here was extraordinary but equally rewarding. The best part about building our runtime using HyperSDK is that we benefit from the optimizations that come to the HyperSDK framework, worked on by some of the best engineers from Ava Labs and the wider open-source HyperSDK community. For example, with the development of HyperSDK Vryx, which recently produced exceptional results with HyperVMs processing over 5 billion transactions at 100,000 transactions per second, once our HelixVM supports Vryx, it will also process 100,000 transactions per second. This is the true hidden power of using HyperSDK to build your own blockchain runtime    Achieving this level of optimization without having to do it ourselves makes it a no-brainer for any team to build their own runtime as well. At Nuklai, we want to build a platform that lets users and businesses monetize their data as easily as building a website, and HyperSDK is the ultimate base layer to achieve that without worrying about how the consensus engine works. We want to build a unique platform that suits our needs, and HyperSDK makes it possible with just a few hundred lines of code. Build the Future of Smart Data With Nuklai Testnet Recently, we released our Nuklai testnet. It only took a few months to get here, and we have many exciting features planned for our HelixVM before the mainnet release, coinciding with HyperSDK’s own roadmap of being production-ready. If you want to get started with our testnet, check out the Nuklai Wallet repository, clone it locally, run it, and start interacting with the testnet. You can request test NAI tokens, create your own tokens, transfer native NAI and other user-created tokens, explore a basic version of the blockchain explorer, and even post something to Nuklai Feed - a simple social media platform. Additionally, the wallet lets you change the Nuklai RPC URL, allowing you to connect to your own Nuklai subnet running locally on your machine. You can watch the demo, and get started today! If you prefer running your own Nuklai subnet locally instead of interacting with the official testnet, you can follow the directions outlined in our documentation. This allows you to do more, such as registering validator nodes for staking, delegating to validators, claiming staking rewards, and more. Welcome to the future where you can move a mountain with your fingers! Written by Kiran Pachhai, Lead Blockchain Architect at Nuklai I remember the day one of my good friends reached out and asked if I had heard about Nuklai . I simply told him, “Not really. That sounds like a topic I would study in my science class.” Fast forward six months, and here I am, building a Layer 1 blockchain network for Nuklai. Who would’ve thought? It is amazing where life takes you in a matter of months. Nuklai Nuklai I will get to the core of this blog about my journey into the world of HyperVMs, but before that, it is essential to know how I even got to this point. The Challenge of High Throughput Soon after I heard about Nuklai, I started to research the project to see what all the hype was about. I was really blown away by Nuklai’s vision to empower every user by allowing them to sell their data to other entities, shifting the power from big businesses to the users themselves. It is a simple yet profound vision because, to achieve it, much had to be done not just at the application level but also at the infrastructure and blockchain protocol levels. When dealing with massive amounts of data daily, the blockchain must handle an immense load that no current platform has achieved yet. People talk about Transactions Per Second (TPS), but few know what it truly takes to achieve such high TPS at the blockchain level. Yes, there are blockchains out there with high throughput, but often, they are very hacky and not true Layer 1s. Usually, the approach is to scale horizontally by creating Layer 2s, Layer 3s, sidechains, and whatnot. These solutions feel like patchwork to me because if you introduce an additional layer to your blockchain, increasing the TPS seems more like an afterthought than something built from the ground up. This is where HyperVMs come into play. HyperVMs, Unlocking High-Performance Blockchains But, what is a HyperVM? It’s essentially a virtual machine (VM) built using the HyperSDK framework . To understand HyperSDK, we first need to understand Avalanche . HyperSDK framework HyperSDK framework Avalanche Avalanche According to their documentation, Avalanche is an open-source platform for building decentralized applications in one interoperable, decentralized, and highly scalable ecosystem. Powered by a uniquely powerful consensus mechanism, Avalanche is the first ecosystem designed to accommodate the scale of global finance, with near-instant transaction finality. Avalanche is an open-source platform for building decentralized applications in one interoperable, decentralized, and highly scalable ecosystem. Powered by a uniquely powerful consensus mechanism, Avalanche is the first ecosystem designed to accommodate the scale of global finance, with near-instant transaction finality. The Avalanche consensus mechanism is ridiculously fast, with less than 2 seconds of finality, meaning your transaction is processed and verified almost instantly. Avalanche also allows you to create your own Layer 1 blockchain (or subnet), and the best part is you are not confined to an Ethereum Virtual Machine (EVM) powered subnet. Despite its large developer community, the EVM is not lightweight and comes with a lot of baggage. On Avalanche, you can also launch your own subnet by developing your own VM in any language as long as it conforms to Avalanche’s Proto interfaces . This is where HyperSDK comes into play. Proto interfaces Proto interfaces Theoretically, you could build your own VM in any language you want, as Avalanche provides that freedom. However, it is also true that it is rather difficult and time-intensive to do from scratch. Even if you use Subnet EVM to get started, it is complex to keep it up-to-date as the core code changes frequently. You need to keep up with maintaining it, which can become resource-intensive. HyperSDK addresses this issue by making it faster, safer, and easier to launch your own optimized blockchain on an Avalanche subnet. Subnet EVM Subnet EVM It hides much of the complexity, allowing you to focus on making your project unique rather than worrying about transaction serialization efficiency or increasing transaction throughput. With HyperSDK, you can build your own blockchain runtime with 500-1000 lines of code and still process thousands of transactions per second without needing a large team of engineers. HelixVM: The 100,000 tx/s Powerhouse? At Nuklai, we built our own blockchain runtime, called ‘HelixVM,’ using the HyperSDK framework with around 200 lines of code in under a week. We added exciting features such as natively issuing tokens, bridging tokens from one subnet to another using Avalanche Warp Messaging (AWM), and developing a staking engine that lets validators register for staking and users delegate their NAI tokens to a validator of their choice to start claiming rewards. With some more additions and custom touches, we added everything with around 500 more lines of code in a couple of months. The journey to get here was extraordinary but equally rewarding. The best part about building our runtime using HyperSDK is that we benefit from the optimizations that come to the HyperSDK framework, worked on by some of the best engineers from Ava Labs and the wider open-source HyperSDK community. For example, with the development of HyperSDK Vryx , which recently produced exceptional results with HyperVMs processing over 5 billion transactions at 100,000 transactions per second, once our HelixVM supports Vryx, it will also process 100,000 transactions per second. This is the true hidden power of using HyperSDK to build your own blockchain runtime Achieving this level of optimization without having to do it ourselves makes it a no-brainer for any team to build their own runtime as well. At Nuklai, we want to build a platform that lets users and businesses monetize their data as easily as building a website, and HyperSDK is the ultimate base layer to achieve that without worrying about how the consensus engine works. HyperSDK Vryx HyperSDK Vryx We want to build a unique platform that suits our needs, and HyperSDK makes it possible with just a few hundred lines of code. Build the Future of Smart Data With Nuklai Testnet Recently, we released our Nuklai testnet . It only took a few months to get here, and we have many exciting features planned for our HelixVM before the mainnet release, coinciding with HyperSDK’s own roadmap of being production-ready. If you want to get started with our testnet, check out the Nuklai Wallet repository, clone it locally, run it, and start interacting with the testnet. Nuklai testnet Nuklai testnet HyperSDK’s own roadmap HyperSDK’s own roadmap Nuklai Wallet Nuklai Wallet You can request test NAI tokens, create your own tokens, transfer native NAI and other user-created tokens, explore a basic version of the blockchain explorer, and even post something to Nuklai Feed - a simple social media platform. Additionally, the wallet lets you change the Nuklai RPC URL, allowing you to connect to your own Nuklai subnet running locally on your machine. You can watch the demo, and get started today! demo , demo If you prefer running your own Nuklai subnet locally instead of interacting with the official testnet, you can follow the directions outlined in our documentation . This allows you to do more, such as registering validator nodes for staking, delegating to validators, claiming staking rewards, and more. documentation documentation Welcome to the future where you can move a mountain with your fingers! Written by Kiran Pachhai, Lead Blockchain Architect at Nuklai

Nuklai Testnet Live: Dive Into HyperVMs, Build on a Scalable Blockchain, and Get Rewarded

নুক্লাই টেস্টনেট লাইভ: হাইপারভিএম-এ ডুব দিন, একটি স্কেলেবল ব্লকচেইন তৈরি করুন এবং পুরস্কৃত করুন

Nuklai Testnet Live: Tauchen Sie ein in HyperVMs, bauen Sie auf einer skalierbaren Blockchain und werden Sie belohnt

Nuklai Testnet Live: sumérgete en HyperVM, construye una cadena de bloques escalable y obtén recompensas

Nuklai Testnet Live : plongez dans les HyperVM, construisez sur une blockchain évolutive et soyez récompensé

नुक्लाई टेस्टनेट लाइव: हाइपरवीएम में गोता लगाएँ, स्केलेबल ब्लॉकचेन बनाएँ, और पुरस्कृत हों

Nuklai テストネット ライブ: HyperVM に飛び込み、スケーラブルなブロックチェーンを構築して報酬を獲得

Nuklai Testnet Live: HyperVM에 뛰어들어 확장 가능한 블록체인을 구축하고 보상을 받으세요

Nuklai Testnet Live: mergulhe nos HyperVMs, desenvolva um blockchain escalável e seja recompensado

Nuklai Testnet Live: погрузитесь в HyperVM, создайте масштабируемый блокчейн и получите вознаграждение

Nuklai Testnet Live: HyperVM'lere Dalın, Ölçeklenebilir Bir Blockchain Üzerine Geliştirin ve Ödül Alın

Nuklai Testnet Live: Đi sâu vào HyperVM, xây dựng trên Blockchain có thể mở rộng và nhận phần thưởng

Nuklai 测试网直播：深入探索 HyperVM，在可扩展的区块链上构建并获得奖励

The Hidden Cost of Hypergrowth: Scale Without Sacrificing Stability

At the SuperAI Conference in Singapore, HyperCycle unveiled HyperRing, a new addition to its line of products aimed at decentralizing AI technology. HyperRing, described as a high-end wearable ring, is designed to protect financial assets generated in the AI economy, merging technology with fashion elements. The HyperRing features a minimalist, battery-less, and sensor-free design, which aims to offer enhanced security compared to existing smart wearables. It utilizes NFC and MFA (multi-factor authentication) to protect digital assets from cyber threats. Toufi Saliba, CEO of HyperCycle, notes that the product is geared towards the growing sector of AI-enhanced wealth, with a focus on combining style with privacy and security features. HyperCycle’s strategy includes the use of advanced protocols such as TODA/IP and Earth64, facilitating what it calls an "Internet of AI". This network is designed to allow AI agents to conduct transactions and learn from each other independently, which could shift the current dynamics in AI interactions. The company’s initiatives also extend to HyperPG and HyperBox, aimed at enhancing AI-to-AI communication with the goal of faster and more effective learning algorithms. HyperCycle's emphasis on a decentralized AI network suggests a move away from traditional cloud computing and AI services towards a more scalable and secure infrastructure, contingent on network performance and user adoption. The introduction of HyperRing at the SuperAI conference reflects HyperCycle’s broader objectives to shape the future of AI-enabled transactions. The integration of decentralized protocols is intended to reduce dependence on centralized computing resources, potentially leading to a more efficient framework for AI interactions. The design of HyperRing, free from batteries and sensors, targets a specific market segment concerned with privacy issues. In today's environment of heightened digital surveillance, this approach could appeal to those seeking secure yet fashionable methods to manage digital assets. Final Thoughts HyperCycle’s approach with HyperRing seems attuned to emerging trends in AI and wearable technology. The product’s security features, combined with its focus on privacy, may appeal to both technologically adept individuals and privacy advocates. As the digital environment evolves, products like HyperRing could influence consumer expectations and set new standards in wearable technology. Success for HyperCycle will likely hinge on its ability to integrate these innovations effectively into everyday user experiences, making the technology both advanced and user-friendly. Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR. At the SuperAI Conference in Singapore, HyperCycle unveiled HyperRing, a new addition to its line of products aimed at decentralizing AI technology. HyperRing, described as a high-end wearable ring, is designed to protect financial assets generated in the AI economy, merging technology with fashion elements. The HyperRing features a minimalist, battery-less, and sensor-free design, which aims to offer enhanced security compared to existing smart wearables. It utilizes NFC and MFA (multi-factor authentication) to protect digital assets from cyber threats. Toufi Saliba, CEO of HyperCycle, notes that the product is geared towards the growing sector of AI-enhanced wealth, with a focus on combining style with privacy and security features. HyperCycle’s strategy includes the use of advanced protocols such as TODA/IP and Earth64, facilitating what it calls an "Internet of AI". This network is designed to allow AI agents to conduct transactions and learn from each other independently, which could shift the current dynamics in AI interactions. The company’s initiatives also extend to HyperPG and HyperBox, aimed at enhancing AI-to-AI communication with the goal of faster and more effective learning algorithms. HyperCycle's emphasis on a decentralized AI network suggests a move away from traditional cloud computing and AI services towards a more scalable and secure infrastructure, contingent on network performance and user adoption. The introduction of HyperRing at the SuperAI conference reflects HyperCycle’s broader objectives to shape the future of AI-enabled transactions. The integration of decentralized protocols is intended to reduce dependence on centralized computing resources, potentially leading to a more efficient framework for AI interactions. The design of HyperRing, free from batteries and sensors, targets a specific market segment concerned with privacy issues. In today's environment of heightened digital surveillance, this approach could appeal to those seeking secure yet fashionable methods to manage digital assets. Final Thoughts Final Thoughts HyperCycle’s approach with HyperRing seems attuned to emerging trends in AI and wearable technology. The product’s security features, combined with its focus on privacy, may appeal to both technologically adept individuals and privacy advocates. As the digital environment evolves, products like HyperRing could influence consumer expectations and set new standards in wearable technology. Success for HyperCycle will likely hinge on its ability to integrate these innovations effectively into everyday user experiences, making the technology both advanced and user-friendly. Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR. Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program . HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR. Vested Interest Disclosure: business blogging program business blogging program

HyperCycle Launches New Security-First Smart Wearable at SuperAI Conference

সুপারএআই কনফারেন্সে হাইপারসাইকেল লঞ্চ করেছে নতুন নিরাপত্তা-প্রথম স্মার্ট পরিধানযোগ্য

HyperCycle stellt auf der SuperAI-Konferenz ein neues, sicherheitsorientiertes Smart Wearable vor

HyperCycle lanza un nuevo dispositivo portátil inteligente que prioriza la seguridad en la conferencia SuperAI

HyperCycle lance un nouveau portable intelligent axé sur la sécurité lors de la conférence SuperAI

हाइपरसाइकल ने सुपरएआई कॉन्फ्रेंस में नया सुरक्षा-प्रथम स्मार्ट वियरेबल लॉन्च किया

HyperCycle が SuperAI カンファレンスでセキュリティ重視の新型スマートウェアラブルを発表

HyperCycle, SuperAI 컨퍼런스에서 새로운 보안 우선 스마트 웨어러블 출시

HyperCycle lança novo wearable inteligente que prioriza a segurança na SuperAI Conference

HyperCycle представляет новое интеллектуальное носимое устройство, ориентированное на безопасность, на конференции SuperAI

HyperCycle, SuperAI Konferansında Güvenliği Ön Plana Çıkaran Yeni Akıllı Giyilebilir Cihazını Piyasaya Sürüyor

HyperCycle ra mắt thiết bị đeo thông minh đầu tiên về bảo mật mới tại Hội nghị SuperAI

HyperCycle 在 SuperAI 大会上推出全新安全第一智能可穿戴设备

Stories, News & Mentions about HYP

Hyp’s Stories on HackerNoon