ABSTRACT
Machine learning models differ in terms of accuracy, computational/memory complexity, training time, and adaptability among other characteristics. 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 for their quick training, computational efficiency, and adaptability. This work presents a hybrid, synergic machine learning model that excels at all the said characteristics and is suitable for incremental, on-line learning on a chip. The proposed model comprises an NN and a classifier. The NN acts as a feature extractor and is specifically trained to work well with the classifier that employs the HD computing framework. This work also presents a parameterized hardware implementation of the said feature extraction and classification components while introducing a compiler that maps any arbitrary NN and/or classifier to the aforementioned hardware. The proposed hybrid machine learning model has the same level of accuracy (i.e. ±1%) as NNs while achieving at least 10% improvement in accuracy compared to HD learning models. Additionally, the end-to-end hardware realization of the hybrid model improves power efficiency by 1.60x compared to state-of-the-art, high-performance HD learning implementations while improving latency by 2.13x. These results have profound implications for the application of such synergic models in challenging cognitive tasks.
- Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In IEEE conference on computer vision and pattern recognition (CVPR), 2016.Google ScholarCross Ref
- Vivienne Sze, Yu-Hsin Chen, Tien-Ju Yang, and Joel S Emer. Efficient processing of deep neural networks: A tutorial and survey. Proceedings of the IEEE, 2017.Google ScholarCross Ref
- Roi Livni, Shai Shalev-Shwartz, and Ohad Shamir. On the computational efficiency of training neural networks. In Advances in neural information processing systems, 2014.Google ScholarDigital Library
- Michael McCloskey and Neal J Cohen. Catastrophic interference in connectionist networks: The sequential learning problem. In Psychology of learning and motivation. Elsevier, 1989.Google Scholar
- Doyen Sahoo, Quang Pham, Jing Lu, and Steven CH Hoi. Online deep learning: Learning deep neural networks on the fly. arXiv preprint arXiv:1711.03705, 2017.Google Scholar
- Viktor Losing, Barbara Hammer, and Heiko Wersing. Incremental on-line learning: A review and comparison of state of the art algorithms. Neurocomputing, 2018.Google Scholar
- Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. ImageNet classification with deep convolutional neural networks. In Advances in neural information processing systems, 2012.Google ScholarDigital Library
- Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, et al. ImageNet large scale visual recognition challenge. International journal of computer vision, 2015.Google ScholarDigital Library
- Pentti Kanerva. Hyperdimensional computing: An introduction to computing in distributed representation with high-dimensional random vectors. Cognitive computation, 2009.Google Scholar
- Mohsen Imani, Chenyu Huang, Deqian Kong, and Tajana Rosing. Hierarchical hyperdimensional computing for energy efficient classification. In ACM/ESDA/IEEE Design Automation Conference (DAC), 2018.Google Scholar
- Justin Morris, Mohsen Imani, Samuel Bosch, Anthony Thomas, Helen Shu, and Tajana Rosing. CompHD: Efficient hyperdimensional computing using model compression. In IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED), 2019.Google ScholarCross Ref
- Jungwook Choi, Zhuo Wang, Swagath Venkataramani, Pierce I-Jen Chuang, Vijayalakshmi Srinivasan, and Kailash Gopalakrishnan. PACT: Parameterized clipping activation for quantized neural networks. arXiv preprint arXiv:1805.06085, 2018.Google Scholar
- Shuchang Zhou, Yuxin Wu, Zekun Ni, Xinyu Zhou, He Wen, and Yuheng Zou. DoReFa-Net: Training low bitwidth convolutional neural networks with low bitwidth gradients. arXiv preprint arXiv:1606.06160, 2016.Google Scholar
- Tianyun Zhang, Shaokai Ye, Kaiqi Zhang, Jian Tang, Wujie Wen, Makan Fardad, and Yanzhi Wang. A systematic DNN weight pruning framework using alternating direction method of multipliers. In European Conference on Computer Vision (ECCV), 2018.Google ScholarCross Ref
- Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531, 2015.Google Scholar
- Manuel Schmuck, Luca Benini, and Abbas Rahimi. Hardware optimizations of dense binary hyperdimensional computing: Rematerialization of hypervectors, binarized bundling, and combinational associative memory. ACM Journal on Emerging Technologies in Computing Systems (JETC), 2019.Google ScholarDigital Library
- Davide Anguita, Alessandro Ghio, Luca Oneto, Xavier Parra, and Jorge Luis Reyes-Ortiz. A public domain dataset for human activity recognition using smartphones. In Esann, 2013.Google Scholar
- Ron Cole, Yeshwant Muthusamy, and Mark Fanty. The ISOLET spoken letter database. Oregon Graduate Institute of Science and Technology, Department of Computer ..., 1990.Google Scholar
- Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Py-Torch: An imperative style, high-performance deep learning library. In Advances in Neural Information Processing Systems, 2019.Google Scholar
- Leslie N Smith and Nicholay Topin. Super-convergence: Very fast training of neural networks using large learning rates. In Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications. International Society for Optics and Photonics, 2019.Google ScholarCross Ref
- Laurens van der Maaten and Geoffrey Hinton. Visualizing data using t-SNE. Journal of machine learning research, 2008.Google Scholar
- Andrey Ignatov. Real-time human activity recognition from accelerometer data using convolutional neural networks. Applied Soft Computing, 2018.Google ScholarCross Ref
- Mohsen Imani, Sahand Salamat, Saransh Gupta, Jiani Huang, and Tajana Rosing. Fach: Fpga-based acceleration of hyperdimensional computing by reducing computational complexity. In Proceedings of the 24th Asia and South Pacific Design Automation Conference, 2019.Google ScholarDigital Library
- Gal Chechik, Uri Shalit, Varun Sharma, and Samy Bengio. An online algorithm for large scale image similarity learning. In Advances in Neural Information Processing Systems, 2009.Google Scholar
- Mohsen Imani, Deqian Kong, Abbas Rahimi, and Tajana Rosing. Voicehd: Hyper-dimensional computing for efficient speech recognition. In 2017 IEEE International Conference on Rebooting Computing (ICRC). IEEE, 2017.Google ScholarCross Ref
- Abbas Rahimi, Pentti Kanerva, and Jan M Rabaey. A robust and energy-efficient classifier using brain-inspired hyperdimensional computing. In Proceedings of the 2016 International Symposium on Low Power Electronics and Design (ISLPED), 2016.Google ScholarDigital Library
- Sohum Datta, Ryan AG Antonio, Aldrin RS Ison, and Jan M Rabaey. A programmable hyper-dimensional processor architecture for human-centric IoT. IEEE Journal on Emerging and Selected Topics in Circuits and Systems, 2019.Google ScholarCross Ref
- Mohsen Imani, Abbas Rahimi, Deqian Kong, Tajana Rosing, and Jan M Rabaey. Exploring hyperdimensional associative memory. In 2017 IEEE International Symposium on High Performance Computer Architecture (HPCA). IEEE, 2017.Google ScholarCross Ref
Index Terms
- SynergicLearning: neural network-based feature extraction for highly-accurate hyperdimensional learning
Recommendations
Control by Learning in a Temperature System Using a Maximum Sensibility Neural Network
MICAI '13: Proceedings of the 2013 12th Mexican International Conference on Artificial IntelligenceA maximum sensibility neural network is implemented in an embedded system to make an online machine learning system, which is used to control the temperature of a small chamber. This is made by manually controlling the temperature to different set-...
ART2 neural network interacting with environment
It is common to train a neural network by using samples so that it can realize the required input-output characteristics. However, to obtain such samples is difficult or even impossible in some cases. This paper proposes the use of on-line reinforcement ...
A scalable pipelined architecture for real-time computation of MLP-BP neural networks
In this paper a novel architecture for implementing multi-layer perceptron (MLP) neural networks on field programmable gate arrays (FPGA) is presented. The architecture presents a new scalable design that allows variable degrees of parallelism in order ...
Comments