线虫生物物理细节电路作为视觉鲁棒机器人操作的任务无关动力学核心

中文摘要 AI 辅助

机器人策略通常针对单一任务、单一本体和单一视觉环境进行训练，并且在这些条件之外泛化能力较差。神经系统能否通过其进化形成的布线和生物物理特性来提供感觉运动计算，这一问题仍未解决。在此，我们将一个具有生物物理细节的秀丽隐杆线虫感觉运动电路——包含136个具有真实形态和电生理特征的多房室神经元——作为视觉运动策略的动力学核心。仅训练薄的任务特定适配器；核心的突触权重保持固定，而其膜电压自由演化。在不同的MetaWorld任务中，该核心匹配或超越了扩散策略、动作分块变换器和神经电路策略基线，并且在视觉扰动下退化更少。用通用网络模型（如MLP、LSTM、Transformer或储备池网络）替换核心会消除这一优势。此外，在真实机器人手臂上，该核心能够承受导致基线崩溃的各种视觉扰动。我们的结果表明，视觉鲁棒性可以从具有生物物理细节的电路动力学中继承，而不是由任务特定控制器学习得到。

英文摘要

Robot policies are usually trained for one task, one body and one visual environment, and generalize poorly beyond these conditions. Whether a nervous system can instead supply the sensorimotor computation through its evolved wiring and biophysics remains unresolved. Here we embed a biophysically detailed Caenorhabditis elegans sensorimotor circuit - 136 multicompartment neurons with realistic morphologies and electrophysiological characteristics - as the dynamical core of a visuomotor policy. Only thin task-specific adapters are trained; the core's synaptic weights stay fixed while its membrane voltages evolve freely. Across different MetaWorld tasks the core matches or exceeds diffusion-policy, action-chunking-transformer and neural-circuit-policy baselines, and degrades less under visual perturbations. Replacing the core with generic network models such as MLP, LSTM, transformer or reservoir networks removes the advantage. Furthermore, on a real robotic arm the core withstands diverse visual perturbations that collapse the baselines. Our results suggest that visual robustness can be inherited from biophysically detailed circuit dynamics rather than learned by task-specific controllers.

发表机构

• CogLeap.AI Space Intelligence (Wuxi) Technology Co., Ltd.（CogLeap.AI 空间智能（无锡）科技有限公司）
• Xiangtan University（湘潭大学）
• Fudan University（复旦大学）
• Tsinghua University（清华大学）

机构由 AI 辅助整理，请以论文原文为准。