Id: arxiv/2611.13291 Url: https://arxiv.org/abs/2611.13291 Tags: eess.SY, math.OC Score: 0 Title: Implicit DeepReach Without Implicit Differentiation Domain: eess.SY Source: arxiv Authors: Xingjian Li, Kelvin Kan, Deepanshu Verma, Krishna Kumar, Hayden Schaeffer, Samy Wu Fung Reviews: 0 Subject: eess.SY Category: research Comments: 0 Disputed: false Division: eess Mentions: 0 Discussed: 0 Published: 2026-11-09 Page: https://delphi.science/r/arxiv/2611.13291 We consider the problem of approximating backward reachable and reach-avoid tubes for high-dimensional nonlinear systems whose Hamiltonians require numerical optimization over constrained controls and disturbances. Existing neural Hamilton-Jacobi (HJ) reachability methods alleviate the dimensionality bottleneck of grid-based solvers, but evaluating Hamiltonians without closed-form expressions introduces an additional computational challenge. We show that, under suitable assumptions, the gradients required for value-function training can be computed without differentiating through the inner…