CoNES: Convex Natural Evolutionary Strategies

Sushant Veer, Anirudha Majumdar

Keywords: blackbox optimization, evolutionary strategies

Abstract: We present a novel algorithm -- convex natural evolutionary strategies (CoNES) -- for optimizing high-dimensional blackbox functions by leveraging tools from convex optimization and information geometry. CoNES is formulated as an efficiently-solvable convex program that adapts the evolutionary strategies (ES) gradient estimate to promote rapid convergence. The resulting algorithm is invariant to the parameterization of the belief distribution. Our numerical results demonstrate that CoNES vastly outperforms conventional blackbox optimization methods on a suite of functions used for benchmarking blackbox optimizers. Furthermore, CoNES demonstrates the ability to converge faster than conventional blackbox methods on a selection of OpenAI's MuJoCo reinforcement learning tasks for locomotion.

One-sentence Summary: We present a novel algorithm -- convex natural evolutionary strategies (CoNES) -- for optimizing high-dimensional blackbox functions by leveraging tools from convex optimization and information geometry.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Community Implementations: [ 2 code implementations](https://www.catalyzex.com/paper/arxiv:2007.08601/code)

Reviewed Version (pdf): https://openreview.net/references/pdf?id=Ogfr3kmOjj