Do Differentiable Simulators Give Better Policy Gradients? — mentioned by blt

Tue, 06 Oct 2026 04:05:12 GMT

Every few years, a derivative-free neural network optimization algorithm gets some hype. I'd bet my life savings that none of them ever make an impact. Derivative-free optimization can be useful for genuinely discontinuous objectives [1], but common neural network objectives are smooth and/or Lipschitz. The gradient is useful. Instead of trying random directions and hoping that one of them is an improvement, it tells you where to go. The more parameters you have, the more useful it becomes. A strict complexity gap between gradient-based and derivative-free Lipschitz convex optimization has…

Paper and discussion history. A mention is not an endorsement.

Do Differentiable Simulators Give Better Policy Gradients? · HackerPapers discussions

Do Differentiable Simulators Give Better Policy Gradients? — mentioned by blt