Brain inspired method solves neural timing learning problem
A Gradient-based yet Spike-Timing-Dependent Solution to the Feedback Learning Problem in Neural Microcircuits
Neural and Evolutionary ComputingMachine Learning
Summary
Figuring out how brain cells learn from the timing of their spikes has been a mystery for decades. The authors propose a new way to think about this learning as separating useful signals in the cells' current state. They develop a method that uses the timing of spikes to assign credit or blame for outcomes, helping neural circuits learn efficiently. Their approach matches or beats other methods while using fewer resources and might explain how real brains learn.
What this means in practice
- •For neuromorphic hardware developers: Design neural chips that learn from spike timing without heavy gradient computations, enabling efficient learning on edge devices.
- •For machine learning engineers: Implement robust spiking neural network models that integrate information over long times and retain memory with fewer parameters.
Authors
Xiangnan Zhang, Jingxin Liu, Ranqi Lu, Jingyu Liu, Qunxi Dong, Fuze Tian, Lixian Zhu, Bin Hu, Björn W. Schuller
Abstract
The brain uses discrete spikes for dynamic computation, yet, how neural microcircuits (NMCs) solve temporal credit assignment using local spike timing remains a fundamental open question. Dominant spiking neural network (SNN) approaches circumvent this by approximating backpropagation through surrogate gradients, decoupling learning from biological spike timing. Here, we reformulate temporal credit assignment as a state separation problem: extracting task-required components induced by historical perturbations directly from the current neural state. This enables an online feedback learning framework for NMCs through a gradient tunneling (GT) algorithm and the lead-lag expansion technique that derives credit assignment from local synaptic spike timing, while remaining compatible with ANN-SNN hybrid architectures. Experimentally, GT-trained NMCs excel at long-timescale evidence integration and noise-robust memory retention, and perform comparably to leading SNN online learning methods on real-world benchmarks with far fewer parameters. The proposed framework addresses the two-decade-old NMC feedback learning problem and suggests a computationally plausible explanation for the brain's learning mechanisms.