Weight transport through spike timing for robust local gradients

2025-03-04

AI Summary

Backpropagation in deep spiking neural networks (SNNs) relies on symmetric weight connections, conflicting with biological locality constraints and hardware implementation requirements. Method: We propose spike-based alignment learning (SAL), a biologically plausible training mechanism grounded in spike-timing statistics. SAL integrates STDP, dual-mode Hebbian/anti-Hebbian plasticity, and intrinsic neuronal noise to adaptively align asymmetric feedforward–feedback weights—without requiring weight symmetry or explicit backward weight transmission—thereby recovering accurate local gradients. The model employs probabilistic spiking neurons and a hierarchical architecture inspired by cortical microcircuits. Contribution/Results: SAL significantly improves convergence accuracy toward target distributions and enables automatic alignment of feedback weights across multiple layers. Local error estimates achieve accuracy comparable to ideal backpropagation. The method satisfies key neurobiological constraints while demonstrating robustness under realistic computational conditions.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Bio-inspired Learning
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: Social networks and social learning

Abstract

In both machine learning and in computational neuroscience, plasticity in functional neural networks is frequently expressed as gradient descent on a cost. Often, this imposes symmetry constraints that are difficult to reconcile with local computation, as is required for biological networks or neuromorphic hardware. For example, wake-sleep learning in networks characterized by Boltzmann distributions builds on the assumption of symmetric connectivity. Similarly, the error backpropagation algorithm is notoriously plagued by the weight transport problem between the representation and the error stream. Existing solutions such as feedback alignment tend to circumvent the problem by deferring to the robustness of these algorithms to weight asymmetry. However, they are known to scale poorly with network size and depth. We introduce spike-based alignment learning (SAL), a complementary learning rule for spiking neural networks, which uses spike timing statistics to extract and correct the asymmetry between effective reciprocal connections. Apart from being spike-based and fully local, our proposed mechanism takes advantage of noise. Based on an interplay between Hebbian and anti-Hebbian plasticity, synapses can thereby recover the true local gradient. This also alleviates discrepancies that arise from neuron and synapse variability -- an omnipresent property of physical neuronal networks. We demonstrate the efficacy of our mechanism using different spiking network models. First, we show how SAL can significantly improve convergence to the target distribution in probabilistic spiking networks as compared to Hebbian plasticity alone. Second, in neuronal hierarchies based on cortical microcircuits, we show how our proposed mechanism effectively enables the alignment of feedback weights to the forward pathway, thus allowing the backpropagation of correct feedback errors.

Problem

• Addresses weight transport problem in neural networks
• Proposes spike-based alignment learning for local gradients
• Improves feedback weight alignment in spiking networks

Innovation

• Spike-based alignment learning (SAL) for neural networks
• Uses spike timing to correct connection asymmetry
• Combines Hebbian and anti-Hebbian plasticity for local gradients