ES is more than just a traditional finite-difference approximator
Open accessThe DOI matches its Crossref record; 4 of the 4 references it cites were found in a registry or an index.
Abstract
An evolution strategy (ES) variant based on a simplification of a natural evolution strategy recently attracted attention because it performs surprisingly well in challenging deep reinforcement learning domains. It searches for neural network parameters by generating perturbations to the current set of parameters, checking their performance, and moving in the aggregate direction of higher reward. Because it resembles a traditional finite-difference approximation of the reward gradient, it can naturally be confused with one. However, this ES optimizes for a different gradient than just reward: It optimizes for the average reward of the entire population, thereby seeking parameters that are robust to perturbation. This difference can channel ES into distinct areas of the search space relative to gradient descent, and also consequently to networks with distinct properties. This unique robustness-seeking property, and its consequences for optimization, are demonstrated in several domains. They include humanoid locomotion, where networks from policy gradient-based reinforcement learning are significantly less robust to parameter perturbation than ES-based policies solving the same task. While the implications of such robustness and robustness-seeking remain open to further study this work's main contribution is to highlight such differences and their potential importance.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Multivariate stochastic approximation using a simultaneous perturbation gradient approximation
Spall, J. C. (1992) · IEEE Transactions on Automatic Control · doi:10.1109/9.119632 · Cited by 2,283 · cited 2×
“Indeed, because of its simplicity there are many policy gradient methods motivated by FD [9, 25].”
BackgroundMethodKirschner, M., & Gerhart, J. (1998) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.95.15.8420 · Cited by 1,129
Robustness and evolvability: a paradox resolved
Wagner, A. (2008) · Proceedings of the Royal Society B: Biological Sciences · doi:10.1098/rspb.2007.1137 · Cited by 669 · cited 3×
“Researchers in EC have long been concerned with robustness in the face of mutation [15, 28, 32], i.”
BackgroundQuality Diversity: A New Frontier for Evolutionary Computation
Pugh, J. K., Soros, L. B., & Stanley, K. O. (2016) · Frontiers in Robotics and AI · doi:10.3389/frobt.2016.00040 · Cited by 514 · cited 1×
“but that again cannot yet be answered. Another open question is how robustness interacts with divergent search techniques like novelty search (Lehman and Stanley, 2011a) or quality diversity methods (Pugh et al., 2016); follow-up experiments to Conti et al. (2017), which combines ES with novelty search, could explore…”
BackgroundSelf-citation
Cited by
Such, F. P., Madhavan, V., Conti, E., Lehman, J., Stanley, K. O., & Clune, J. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1712.06567 · Cited by 525
Self-citationClune, J. (2019) · arXiv · doi:10.48550/arxiv.1905.10985 · Cited by 82
“There are pros and cons to using evolutionary methods [176, 96, 146, 195, 167], and benefits to hybrid approaches that combine evolutionary concepts (searching in the parameter space) with policy-gradient concepts (searching in the action space) [131, 46].”
BackgroundSelf-citationPOET: open-ended coevolution of environments and their optimized solutions
Wang, R., Lehman, J., Clune, J., & Stanley, K. O. (2019) · Proceedings of the Genetic and Evolutionary Computation Conference · doi:10.1145/3321707.3321799 · Cited by 61
Self-citationLehman, J. (2019) · arXiv · doi:10.48550/arxiv.1906.10189 · Cited by 3
Self-citation
Metadata
- Authors
- Lehman, J., Chen, J., Clune, J., & Stanley, K. O.
- Venue
- Proceedings of the Genetic and Evolutionary Computation Conference
- Publisher
- ACM
- Published
- 2018-07-02
- Pages
- 450-457
- OpenAlex
- W2776365419
- Cited by
- 63
- References
- 4
- FWCI
- 4.68
- Type
- conference paper
- Language
- en
- Open access
- bronze
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.