Filters








183 Hits in 3.5 sec

Adaptive neural recovery for highly robust brain-like representation

Prathyush Poduval, Yang Ni, Yeseong Kim, Kai Ni, Raghavan Kumar, Rossario Cammarota, Mohsen Imani
2022 Proceedings of the 59th ACM/IEEE Design Automation Conference  
In this paper, we propose RobustHD, a robust and noise-tolerant learning system based on HyperDimensional Computing (HDC), mimicking important brain functionalities.  ...  Unlike traditional binary representation, RobustHD exploits a redundant and holographic representation, ensuring all bits have the same impact on the computation.  ...  The results are reported for a human activity recognition task (UCIHAR) [20] using hypervectors with different dimensions (𝐷 = 5𝑘 and 𝐷 = 10𝑘) and different bit precision.  ... 
doi:10.1145/3489517.3530659 fatcat:twse7qawbzhk7lf7j7owrfyyde

Flexible and Personalized Learning for Wearable Health Applications using HyperDimensional Computing

Sina Shahhosseini, Yang Ni, Emad Kasaeyan Naeini, Mohsen Imani, Amir M. Rahmani, Nikil Dutt
2022 Proceedings of the Great Lakes Symposium on VLSI 2022  
We exploit a Hyperdimensional computing (HDC) solution for wearable devices that offers flexibility, high efficiency, and performance while enabling on-device personalization and privacy protection.  ...  CCS CONCEPTS • Computer systems organization → Embedded software.  ...  We evaluate our proposed system using three health case studies: Pain Monitoring (PM), Stress Monitoring (SM), Human Activity Recognition (HAR).  ... 
doi:10.1145/3526241.3530373 fatcat:kzwh3km4dbfcphtoje3cohf4sa

Spiking Hyperdimensional Network: Neuromorphic Models Integrated with Memory-Inspired Framework [article]

Zhuowen Zou, Haleh Alimohamadi, Farhad Imani, Yeseong Kim, Mohsen Imani
2021 arXiv   pre-print
Particularly, Spiking Neural Networks (SNNs) and HyperDimensional Computing (HDC) have shown promising results in enabling efficient and robust cognitive learning.  ...  Recently, brain-inspired computing models have shown great potential to outperform today's deep learning solutions in terms of robustness and energy efficiency.  ...  In the HDC, binary (bipolar) hypervectors are often used for computation efficiency.  ... 
arXiv:2110.00214v1 fatcat:kjaxx2xjzze7zbbl56h2bbwemu

EnHDC: Ensemble Learning for Brain-Inspired Hyperdimensional Computing [article]

Ruixuan Wang, Dongning Ma, Xun Jiao
2022 arXiv   pre-print
Recently, brain-inspired hyperdimensional computing (HDC) becomes an emerging computational paradigm that has achieved success in various domains such as human activity recognition, voice recognition,  ...  EnHDC uses a majority voting-based mechanism to synergistically integrate the prediction outcomes of multiple base HDC classifiers.  ...  Experiment Setup We evaluate EnHDC using four application domains: speech recognition (ISOLET [5] ), human activity recognition (HAR [1] ), handwritten digits (MNIST [14] ), and cardiotocography (CARDIO  ... 
arXiv:2203.13542v1 fatcat:k6ikdv6gcbeehjti6qgh7nhsb4

Energy Efficient In-memory Hyperdimensional Encoding for Spatio-temporal Signal Processing

Geethan Karunaratne, Manuel Le Gallo, Michael Hersche, Giovanni Cherubini, Luca Benini, Abu Sebastian, Abbas Rahimi
2021 IEEE Transactions on Circuits and Systems - II - Express Briefs  
The emerging brain-inspired computing paradigm known as hyperdimensional computing (HDC) has been proven to provide a lightweight learning framework for various cognitive tasks compared to the widely used  ...  In particular, we introduce a methodology for the in-memory hyperdimensional encoding of ST data to be used together with an in-memory associative search module.  ...  Hyperdimensional computing (HDC) [1] is one promising brain-inspired computing approach that relies on representing entities using high-dimensional (up to 10,000 dimensions) vectors called hypervectors  ... 
doi:10.1109/tcsii.2021.3068126 fatcat:67lzv4tudje4lkguzce6mtqsbu

Hyperdimensional Computing-based Multimodality Emotion Recognition with Physiological Signals

En-Jui Chang, Abbas Rahimi, Luca Benini, An-Yeu Andy Wu
2019 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS)  
To overcome this issue, brain-inspired hyperdimensional (HD) computing, an energy-efficient and fast learning computational paradigm, has a high potential to achieve a balance between accuracy and the  ...  To interact naturally and achieve mutual sympathy between humans and machines, emotion recognition is one of the most important function to realize advanced human-computer interaction devices.  ...  Background of HD Computing The operation of the human brain relies on billions of neurons and synapses, suggesting that massive neural activities are fundamental to its computational power.  ... 
doi:10.1109/aicas.2019.8771622 dblp:conf/aicas/ChangRBW19 fatcat:brsbrxz3prdtlenhgpemey4vnu

QHD: A brain-inspired hyperdimensional reinforcement learning algorithm [article]

Yang Ni, Danny Abraham, Mariam Issa, Yeseong Kim, Pietro Mercati, Mohsen Imani
2022 arXiv   pre-print
., Deep Q-Learning, are based on deep neural networks, putting high computational costs when running on edge devices.  ...  This makes QHD suitable for highly-efficient reinforcement learning in the edge environment, where it is crucial to support online and real-time learning.  ...  Hyperdimensional Computing: Hyperdimensional Computing is a brain-like computational model and an alternative lightweight machine learning algorithm.  ... 
arXiv:2205.06978v2 fatcat:pic6uubyync4rnnhg2ry6i7wsy

Classification using Hyperdimensional Computing: A Review [article]

Lulu Ge, Keshab K. Parhi
2020 arXiv   pre-print
Evaluations indicate that HD computing shows great potential in addressing problems using data in the form of letters, signals and images.  ...  Hyperdimensional (HD) computing is built upon its unique data type referred to as hypervectors. The dimension of these hypervectors is typically in the range of tens of thousands.  ...  Index Terms-Hyperdimensional (HD) computing, classification accuracy, energy efficiency. I.  ... 
arXiv:2004.11204v1 fatcat:ttk6u2fi7jerpjxgesyh6bin34

Hyperdimensional Computing in Industrial Systems: The Use-Case of Distributed Fault Isolation in a Power Plant

Denis Kleyko, Evgeny Osipov, Nikolaos Papakonstantinou, Valeriy Vyatkin
2018 IEEE Access  
The proposed approach is based on the principles of hyperdimensional computing. In particular, the recently proposed method called Holographic Graph Neuron is used.  ...  INDEX TERMS Vector symbolic architectures, Holographic Graph Neuron, distributed representation, complex systems, distributed fault isolation, hyperdimensional computing, machine learning. 30766 2169-3536  ...  The proposed methods were exemplified via applications in the following areas: a human activity recognition using the accelerometer data and predictions (e.g., the next app to be loaded) using real-life  ... 
doi:10.1109/access.2018.2840128 fatcat:ew37xuv6pfcctitlqyoeo7oiey

L3E-HD: A Framework Enabling Efficient Ensemble in High-Dimensional Space for Language Tasks

Fangxin Liu, Haomin Li, Xiaokang Yang, Li Jiang
2022 Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval  
Brain-inspired hyperdimensional computing (HDC) has been introduced as an alternative computing paradigm to achieve efficient and robust learning.  ...  Language tasks, generally solved using machine learning methods, are widely deployed on low-power embedded devices.  ...  HDC represents data in the high-dimensional space to mimic the behaviors of the brain computed with patterns of neural activity [14] .  ... 
doi:10.1145/3477495.3531761 fatcat:fvpa2jlwyrhezkrelneeudnbta

Compressing Subject-specific Brain-Computer Interface Models into One Model by Superposition in Hyperdimensional Space

Michael Hersche, Philipp Rupp, Luca Benini, Abbas Rahimi
2020 2020 Design, Automation & Test in Europe Conference & Exhibition (DATE)  
Our method makes the use of unexploited capacity of trained models by orthogonalizing parameters in a hyperdimensional space, followed by iterative retraining to compensate noisy decomposition.  ...  Accurate multiclass classification of electroencephalography (EEG) signals is still a challenging task towards the development of reliable motor imagery brain-computer interfaces (MI-BCIs).  ...  INTRODUCTION Brain-computer interfaces (BCIs) aim to provide a communication and control channel based on the recognition of the subjects intentions from neural activity typically recorded by noninvasive  ... 
doi:10.23919/date48585.2020.9116447 dblp:conf/date/HerscheRBR20 fatcat:hhox3kw2bfh4jclyapdqixitse

Efficient Personalized Learning for Wearable Health Applications using HyperDimensional Computing [article]

Sina Shahhosseini, Yang Ni, Hamidreza Alikhani, Emad Kasaeyan Naeini, Mohsen Imani, Nikil Dutt, Amir M. Rahmani
2022 arXiv   pre-print
Hyperdimensional computing (HDC) offers a well-suited on-device learning solution for resource-constrained devices and provides support for privacy-preserving personalization.  ...  We evaluate the efficacy of our approach using three case studies and show that our system improves the energy efficiency of training by up to 45.8× compared with the state-of-the-art Deep Neural Network  ...  We demonstrated the efficacy of our HDC approach using three realistic wearable healthcare studies, achieving better energy efficiency for training and inference by up to 45.8× and 5.1× compared to state-of-the-art  ... 
arXiv:2208.01095v1 fatcat:odm7btur4bafzm6c47ylfma7ii

Connectionist-Symbolic Machine Intelligence using Cellular Automata based Reservoir-Hyperdimensional Computing [article]

Ozgur Yilmaz
2015 arXiv   pre-print
Also, binary reservoir feature vectors can be combined using Boolean operations as in hyperdimensional computing, paving a direct way for concept building and symbolic processing.  ...  It is possible to estimate the kernel for linear cellular automata via metric learning, that enables a much more efficient distance computation in support vector machine framework.  ...  We conjecture that this retrieval is the essence of recurrent computation in human cortex. Using the power of hyperdimensional computing we can use labels, attributes and predicates to achieve: 1.  ... 
arXiv:1503.00851v3 fatcat:b4mbomhexjezfitvhsny4yjkga

SynergicLearning: Neural Network-Based Feature Extraction for Highly-Accurate Hyperdimensional Learning [article]

Mahdi Nazemi, Amirhossein Esmaili, Arash Fayyazi, Massoud Pedram
2020 arXiv   pre-print
for their quick training, computational efficiency, and adaptability.  ...  For example, neural networks (NNs) are well-known for their high accuracy due to the quality of their automatic feature extraction while brain-inspired hyperdimensional (HD) learning models are famous  ...  To study the effectiveness of SynergicLearning, we use two publicly available datasets: Human Activity Recognition (HAR) [17] and ISOLET [18] .  ... 
arXiv:2007.15222v2 fatcat:rq3u2qx5rvaoraswciwkwkx6eq

MIMHD: Accurate and Efficient Hyperdimensional Inference Using Multi-Bit In-Memory Computing [article]

Arman Kazemi, Mohammad Mehdi Sharifi, Zhuowen Zou, Michael Niemier, X. Sharon Hu, Mohsen Imani
2021 arXiv   pre-print
Hyperdimensional Computing (HDC) is an emerging computational framework that mimics important brain functions by operating over high-dimensional vectors, called hypervectors (HVs).  ...  Thus, we propose a highly accurate and efficient multi-bit in-memory HDC inference platform called MIMHD.  ...  [25], [26] UCIHAR 561 12 6,213 1,554 Activity Recognition(Mobile) [27] ISOLET 617 26 6,238 1,559 Voice Recognition [28] PAMAP 75 5 611,142 101,582 Activity Recognition(IMU) [29] FACE  ... 
arXiv:2106.12029v1 fatcat:npmwowsm2zhzpe24ue63crhucm
« Previous Showing results 1 — 15 out of 183 results