Hyperlyve:并行枚举为深度 ReLU 网络做精确 Lyapunov 稳定性认证
Hyperlyve: Parallelized Enumeration and Exact Lyapunov Stability Certification for Deep ReLU Networks
做神经网络控制器形式化验证的论文值得一试:用并行枚举 ReLU 超平面排布把速度拉快 150 倍,还能对精确动力学做可靠且完备的稳定性证明,比现有 SOTA 验证器多解出一批此前算不动的系统,搞控制安全的朋友可以跟进。
Hyperlyve 是一个认证非线性动力系统渐近稳定性的框架,能处理含神经网络控制器的系统,学习并正式认证连续分段仿射(CPA)神经 Lyapunov 函数。
它在反例引导归纳合成(CEGIS)循环中工作,核心创新是并行算法:枚举由深层网络 ReLU 激活形成的超平面排布,并在每个多面体区域内验证 Lyapunov 条件成立。论文推导了可靠且完备的区域测试方法,直接对精确光滑动力学(而非分段线性近似)在紧致域内验证渐近稳定性,并将 CPA 神经 Lyapunov 验证问题的计算复杂度形式化刻画为 co-NP 完全,填补了文献空白。
实证结果显示,Hyperlyve 枚举 ReLU 网络的速度超过同类工具 150 倍,因此能证明一些对最先进验证器来说目前不可解的系统的渐近稳定性。
Hyperlyve: Parallelized Enumeration and Exact Lyapunov Stability Certification for Deep ReLU Networks
We introduce Hyperlyve (Hyperplane Partitioned Lyapunov Verification and Enumeration), a framework for proving the asymptotic stability of nonlinear dynamical systems, including those with neural network controllers. Hyperlyve learns and formally certifies continuous piecewise-affine (CPA) neural Lyapunov functions within a counterexample-guided inductive synthesis (CEGIS) loop. Hyperlyve's key innovation is a parallel algorithm that enumerates the hyperplane arrangement formed by the ReLU activations of deep networks and certifies that Lyapunov conditions hold within each polyhedral region. We derive sound and complete regional tests to verify the asymptotic stability of the exact smooth dynamics (rather than a piecewise-linear approximation) within a compact domain of interest. Furthermore, we formally characterize the computational complexity of the CPA neural Lyapunov verification problem as co-NP-complete, filling a key gap in the literature. Empirical results demonstrate that Hyperlyve enumerates ReLU networks over 150-times faster than competing tools, which contributes to Hyperlyve proving the asymptotic stability of systems that are currently intractable for state-of-the-art verifiers.
本条情报归属 AITOP「论文」核心追踪栏目。在跨区域产业观察坐标中,该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线,反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括:Lyapunov函数、形式化验证、ReLU网络、神经网络控制器,读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明:本文原始线索采编自「arXiv cs.LG」(@Brian Yarbrough, Jesse Wayment, Jingbo Wang, Philip E. Paré, Shreyas Sundaram),经 AITOP 多维评分管线结构化提炼。