Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
162 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Unleashing the Potential of Spiking Neural Networks by Dynamic Confidence (2303.10276v3)

Published 17 Mar 2023 in cs.CV and cs.NE

Abstract: This paper presents a new methodology to alleviate the fundamental trade-off between accuracy and latency in spiking neural networks (SNNs). The approach involves decoding confidence information over time from the SNN outputs and using it to develop a decision-making agent that can dynamically determine when to terminate each inference. The proposed method, Dynamic Confidence, provides several significant benefits to SNNs. 1. It can effectively optimize latency dynamically at runtime, setting it apart from many existing low-latency SNN algorithms. Our experiments on CIFAR-10 and ImageNet datasets have demonstrated an average 40% speedup across eight different settings after applying Dynamic Confidence. 2. The decision-making agent in Dynamic Confidence is straightforward to construct and highly robust in parameter space, making it extremely easy to implement. 3. The proposed method enables visualizing the potential of any given SNN, which sets a target for current SNNs to approach. For instance, if an SNN can terminate at the most appropriate time point for each input sample, a ResNet-50 SNN can achieve an accuracy as high as 82.47% on ImageNet within just 4.71 time steps on average. Unlocking the potential of SNNs needs a highly-reliable decision-making agent to be constructed and fed with a high-quality estimation of ground truth. In this regard, Dynamic Confidence represents a meaningful step toward realizing the potential of SNNs.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (54)
  1. Dynamic capacity networks. In International Conference on Machine Learning, pages 2549–2558. PMLR, 2016.
  2. A solution to the learning dilemma for recurrent networks of spiking neurons. Nature communications, 11(1):1–15, 2020.
  3. Optimal ann-snn conversion for high-accuracy and ultra-low-latency spiking neural networks. In International Conference on Learning Representations, 2021.
  4. Dynamic convolution: Attention over convolution kernels. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11030–11039, 2020.
  5. Binaryconnect: Training deep neural networks with binary weights during propagations. Advances in neural information processing systems, 28, 2015.
  6. Are SNNs Really More Energy-Efficient Than ANNs? an In-Depth Hardware-Aware Study. IEEE Transactions on Emerging Topics in Computational Intelligence, pages 1–11, 2022.
  7. Comparison of artificial and spiking neural networks on digital hardware. Frontiers in Neuroscience, 15:651141, 2021.
  8. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009.
  9. Optimal conversion of conventional artificial neural networks to spiking neural networks. arXiv preprint arXiv:2103.00476, 2021.
  10. Fast-classifying, high-accuracy spiking deep networks through weight and threshold balancing. In 2015 International Joint Conference on Neural Networks (IJCNN), pages 1–8. ieee, 2015.
  11. Snn-rat: Robustness-enhanced spiking neural network through regularized adversarial training. Advances in Neural Information Processing Systems, 35:24780–24793, 2022.
  12. Optimal ann-snn conversion for fast and accurate inference in deep spiking neural networks. arXiv preprint arXiv:2105.11654, 2021.
  13. Accelerating training of deep spiking neural networks with parameter initialization. 2021.
  14. Neural architecture search: A survey. The Journal of Machine Learning Research, 20(1):1997–2017, 2019.
  15. Learned step size quantization. arXiv preprint arXiv:1902.08153, 2019.
  16. Tom Fawcett. An introduction to roc analysis. Pattern recognition letters, 27(8):861–874, 2006.
  17. Dynamic channel pruning: Feature boosting and suppression. arXiv preprint arXiv:1810.05331, 2018.
  18. The evolution of robotics research. IEEE Robotics & Automation Magazine, 14(1):90–103, 2007.
  19. On calibration of modern neural networks. In International conference on machine learning, pages 1321–1330. PMLR, 2017.
  20. Song Han. Efficient methods and hardware for deep learning. PhD thesis, Stanford University, 2017.
  21. Deep compression: Compressing deep neural networks with pruning, trained quantization and huffman coding. arXiv preprint arXiv:1510.00149, 2015.
  22. Reducing ann-snn conversion error through residual membrane potential. arXiv preprint arXiv:2302.02091, 2023.
  23. Bridging the gap between anns and snns by calibrating offset spikes. arXiv preprint arXiv:2302.10685, 2023.
  24. Channel pruning for accelerating very deep neural networks. In Proceedings of the IEEE international conference on computer vision, pages 1389–1397, 2017.
  25. Mobilenets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:1704.04861, 2017.
  26. Channel gating neural networks. Advances in Neural Information Processing Systems, 32, 2019.
  27. Multi-scale dense networks for resource efficient image classification. arXiv preprint arXiv:1703.09844, 2017.
  28. Densely connected convolutional networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 4700–4708, 2017.
  29. Structured Dynamic Precision for Deep Neural Networks Quantization. ACM Transactions on Design Automation of Electronic Systems, page 3549535, July 2022.
  30. Squeezenet: Alexnet-level accuracy with 50x fewer parameters and¡ 0.5 mb model size. arXiv preprint arXiv:1602.07360, 2016.
  31. An FPGA Implementation of Deep Spiking Neural Networks for Low-Power and Fast Classification. Neural Computation, 32(1):182–204, 2020.
  32. Spiking-YOLO: Spiking Neural Network for Energy-Efficient Object Detection. Proceedings of the AAAI Conference on Artificial Intelligence, 34(07):11270–11277, 2020.
  33. Neural architecture search for spiking neural networks. In Computer Vision–ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23–27, 2022, Proceedings, Part XXIV, pages 36–56. Springer, 2022.
  34. Learning multiple layers of features from tiny images. 2009.
  35. Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25, 2012.
  36. Towards biologically-plausible neuron models and firing rates in high-performance deep spiking neural networks. In International Conference on Neuromorphic Systems 2021, pages 1–7, 2021.
  37. Quantization framework for fast spiking neural networks. Frontiers in Neuroscience, page 1055, 2022.
  38. A free lunch from ann: Towards efficient, accurate spiking neural networks calibration. arXiv preprint arXiv:2106.06984, 2021.
  39. Converting artificial neural networks to spiking neural networks via parameter calibration. arXiv preprint arXiv:2205.10121, 2022.
  40. Runtime neural pruning. Advances in neural information processing systems, 30, 2017.
  41. On-device training under 256kb memory. arXiv preprint arXiv:2206.15472, 2022.
  42. Progressive neural architecture search. In Proceedings of the European conference on computer vision (ECCV), pages 19–34, 2018.
  43. Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3431–3440, 2015.
  44. Wolfgang Maass. Networks of spiking neurons: the third generation of neural network models. Neural networks, 10(9):1659–1671, 1997.
  45. Event-driven random back-propagation: Enabling neuromorphic deep learning machines. Frontiers in neuroscience, 11:324, 2017.
  46. Can you trust your model’s uncertainty? evaluating predictive uncertainty under dataset shift. Advances in neural information processing systems, 32, 2019.
  47. Neural population coding for effective temporal classification. In 2019 International Joint Conference on Neural Networks (IJCNN), pages 1–8. IEEE, 2019.
  48. T2fsnn: Deep spiking neural networks with time-to-first-spike coding. In 2020 57th ACM/IEEE Design Automation Conference (DAC), pages 1–6. IEEE, 2020.
  49. Deep learning with spiking neurons: opportunities and challenges. Frontiers in neuroscience, page 774, 2018.
  50. You only look once: Unified, real-time object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 779–788, 2016.
  51. Conversion of continuous-valued deep networks to efficient event-driven networks for image classification. Frontiers in neuroscience, 11:682, 2017.
  52. DRQ: Dynamic Region-based Quantization for Deep Neural Network Acceleration. In 2020 ACM/IEEE 47th Annual International Symposium on Computer Architecture (ISCA), pages 1010–1021, May 2020.
  53. Direct training for spiking neural networks: Faster, larger, better. In Proceedings of the AAAI conference on artificial intelligence, volume 33, pages 1311–1318, 2019.
  54. Blockdrop: Dynamic inference paths in residual networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 8817–8826, 2018.
Citations (14)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now