Papers

Neural ODEs与并发学习结合实现稳定在线学习

Neural ODEs Meet Concurrent Learning: Stable Online Learning with Lyapunov Guarantees

Why we picked it

Neural ODEs首次获得Lyapunov稳定性保证,在动态系统学习中表现优异,尤其在噪声环境下优势明显。

研究人员证明了Neural ODEs的伴随梯度包含正半定轨迹算子作用于参数误差的结构,为在线学习提供了Lyapunov保证。

Body

该方法在四个DeepMind Control Suite领域测试,在速度测量噪声下,三个领域达到最低中位预测误差,比基于观察器的并发学习性能提升最多8倍。在清洁测量条件下,在pendulum任务上表现最佳,在cartpole和reacher任务上与最佳基线方法相差不超过1.

6倍。

Original · arXiv: Google DeepMind

Neural ODEs Meet Concurrent Learning: Stable Online Learning with Lyapunov Guarantees

Neural ODEs learn dynamics from trajectory losses, but their adjoint gradients lack the regressor-times-parameter-error structure on which Lyapunov analyses of online adaptation rest, so training on streaming data comes without stability guarantees. We show that this structure is in fact present: the adjoint gradient decomposes exactly into a positive semi-definite trajectory operator acting on the parameter error plus a nonlinear perturbation with explicit, horizon-dependent bounds. A quadratic Lyapunov function then certifies online Neural ODE training over sliding windows under computable gain and horizon conditions, and the same certificate extends to stored data: its drift branch recovers concurrent learning, and its trajectory branch yields NODE-CL, a stored-segment Gauss-Newton method built on batched forward sensitivities that needs no state-derivative estimates. On four DeepMind Control Suite domains, NODE-CL attains the lowest median prediction error on three under velocity measurement noise, where observer-based concurrent learning degrades by up to 8x; with clean measurements it is best on the pendulum and within a factor of 1.6 of the best stored-data baseline on the cartpole and reacher.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Papers" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Neural ODEs, Lyapunov, online learning, DeepMind Control Suite.

Primary source verification: published directly by arXiv: Google DeepMind with first-party authority.