← 研踪 PaperTrailarXiv:2605.19403 · v2
TIDE: Asymmetric Neural Circuits for Stabilized Temporal Inhibitory-Excitatory Dynamics
首次提交 2026-05-19 13:59
版本更新 2026-11-16 06:58
雷达观测 2026-11-17 17:12
摘要
Recent Continuous Thought Machine (CTM) architecture decouples internal computation from external inputs via neural dynamics, but relies on unconstrained multi-layer perceptrons for its recurrent computations. We propose to model neural dynamics using asymmetric Excitatory-Inhibitory (E-I) networks, which can be stabilized via principles from network theory, while obeying Dale's principle and can be expressed as energy-based systems optimized through a game-theoretic loss. Building on this perspective, we introduce Temporal Inhibitory-Excitatory Dynamic Engine (TIDE), a neuro-inspired architecture that computes internal representations through neural dynamics stabilized by incorporating the Wilson-Cowan dynamics and lateral inhibition. TIDE balances biological realism by, for instance, using Hierarchical Receptive Fields and enforcing Dale's principle to ensure a realistic E-I balance ratio with an end-to-end trainable architecture. The aim of this paper is to introduce a new architecture that brings neuro-inspired learning and architectural principles to the forefront. We present proofs of convergence, stability, and complexity bounds, along with empirical ablation studies. Overall, TIDE achieves higher accuracy than CTM on the majority of the investigated datasets while requiring a smaller training budget. Furthermore, across varying training budgets, TIDE improves accuracy on ImageNet-C by an average of % to % under different perturbation severity levels.
作者备注:Accepted in the 40th Conference on Neural Information Processing Systems [NeurIPS 2026]
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v2 · 2026-11-16 06:58 · 最新观测版本
TIDE: Asymmetric Neural Circuits for Stabilized Temporal Inhibitory-Excitatory Dynamics
Recent Continuous Thought Machine (CTM) architecture decouples internal computation from external inputs via neural dynamics, but relies on unconstrained multi-layer perceptrons for its recurrent computations. We propose to model neural dynamics using asymmetric Excitatory-Inhibitory (E-I) networks, which can be stabilized via principles from network theory, while obeying Dale's principle and can be expressed as energy-based systems optimized through a game-theoretic loss. Building on this perspective, we introduce Temporal Inhibitory-Excitatory Dynamic Engine (TIDE), a neuro-inspired architecture that computes internal representations through neural dynamics stabilized by incorporating the Wilson-Cowan dynamics and lateral inhibition. TIDE balances biological realism by, for instance, using Hierarchical Receptive Fields and enforcing Dale's principle to ensure a realistic E-I balance ratio with an end-to-end trainable architecture. The aim of this paper is to introduce a new architecture that brings neuro-inspired learning and architectural principles to the forefront. We present proofs of convergence, stability, and complexity bounds, along with empirical ablation studies. Overall, TIDE achieves higher accuracy than CTM on the majority of the investigated datasets while requiring a smaller training budget. Furthermore, across varying training budgets, TIDE improves accuracy on ImageNet-C by an average of % to % under different perturbation severity levels.
查看此版本解读 →查看此版本原文 ↗