[论文解读] Event-based Backpropagation for Analog Neuromorphic Hardware

Christian Pehle, Luca Blessing+3 more|arXiv (Cornell University)|Feb 13, 2023

Advanced Memory and Neural Computing被引用 4

一句话总结

本文提出EventProp，一种用于模拟脉冲神经形态硬件的事件基反向传播算法，仅利用脉冲时间观测来估计梯度，在信息效率方面相比替代梯度方法提升十倍。该方法实现了大规模神经形态系统中可扩展、低能耗、设备端学习，且无需密集的状态测量。

ABSTRACT

Neuromorphic computing aims to incorporate lessons from studying biological nervous systems in the design of computer architectures. While existing approaches have successfully implemented aspects of those computational principles, such as sparse spike-based computation, event-based scalable learning has remained an elusive goal in large-scale systems. However, only then the potential energy-efficiency advantages of neuromorphic systems relative to other hardware architectures can be realized during learning. We present our progress implementing the EventProp algorithm using the example of the BrainScaleS-2 analog neuromorphic hardware. Previous gradient-based approaches to learning used "surrogate gradients" and dense sampling of observables or were limited by assumptions on the underlying dynamics and loss functions. In contrast, our approach only needs spike time observations from the system while being able to incorporate other system observables, such as membrane voltage measurements, in a principled way. This leads to a one-order-of-magnitude improvement in the information efficiency of the gradient estimate, which would directly translate to corresponding energy efficiency improvements in an optimized hardware implementation. We present the theoretical framework for estimating gradients and results verifying the correctness of the estimation, as well as results on a low-dimensional classification task using the BrainScaleS-2 system. Building on this work has the potential to enable scalable gradient estimation in large-scale neuromorphic hardware as a continuous measurement of the system state would be prohibitive and energy-inefficient in such instances. It also suggests the feasibility of a full on-device implementation of the algorithm that would enable scalable, energy-efficient, event-based learning in large-scale analog neuromorphic hardware.

研究动机与目标

• 在大规模模拟神经形态硬件中实现可扩展、节能、设备端的基于梯度的学习。
• 克服先前梯度估计方法的局限性，这些方法需要对神经元动态进行密集采样，或依赖于对网络动态或拓扑的限制性假设。
• 证明仅凭脉冲时间观测即可实现神经形态硬件中准确的梯度估计。
• 验证该方法在硬件可变性和非理想性下的鲁棒性。
• 仅使用事件基数据，实现在物理神经形态芯片上的完整闭环训练。

提出的方法

• EventProp算法仅使用神经形态硬件中记录的脉冲时间观测来估计梯度，避免了对连续状态监控的需求。
• 其在PyTorch中采用基于软件的伴随动力学模型，从记录的脉冲事件和可选的膜电位轨迹中计算参数梯度。
• 该方法通过离散欧拉积分对漏电积分-发放（LIF）神经元动力学进行前向与反向传播的整合，包含脉冲触发的跳跃。
• 梯度通过伴随法计算，损失函数基于最大时间输出电压的公式化设计，用于分类任务。
• 该算法设计为兼容任意网络拓扑、动力学和损失函数，无需对时间常数或放电行为做出假设。
• 在BrainScaleS-2平台上执行软硬件协同训练，其中发送输入脉冲，记录输出脉冲和电压，并在软件中估计梯度。

实验结果

研究问题

• RQ1 是否仅使用脉冲时间观测，无需密集状态测量，即可在模拟神经形态硬件中实现梯度估计？
• RQ2 EventProp算法是否在学习性能上与替代梯度方法相当，同时显著提升信息效率？
• RQ3 该梯度估计在真实神经形态基底上的硬件可变性和非理想性下是否具有鲁棒性？
• RQ4 该算法是否可成功应用于物理神经形态芯片，实现端到端闭环训练？
• RQ5 在硬件环境中，EventProp相比传统替代梯度方法的信息效率增益是多少？

主要发现

• EventProp算法在Yin-Yang数据集上的分类性能与替代梯度方法相当，无论在仿真中还是在BrainScaleS-2硬件上均表现一致。
• 该方法将所需数据量减少了一个数量级，证明了信息效率提升十倍。
• 基于硬件脉冲时间测量的平均梯度估计值与解析解高度吻合，证实了其对基底可变性的鲁棒性。
• 该算法成功实现了仅使用事件基观测数据，在物理模拟神经形态硬件上进行闭环训练。
• 即使仅使用脉冲时间，不依赖连续膜电位记录，梯度估计仍保持高精度。
• 该方法兼容任意网络拓扑、动力学和损失函数，克服了现有方法的先前局限性。