{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/actor-critic-versus-direct-policy-search-a","title":"Actor-critic versus direct policy search: a comparison based on sample complexity","arxiv_id":"1606.09152","date":"2016-06-29","proceeding":null,"authors":["Arnaud de Froissard de Broissia","Olivier Sigaud"],"abstract":"Sample efficiency is a critical property when optimizing policy parameters\nfor the controller of a robot. In this paper, we evaluate two state-of-the-art\npolicy optimization algorithms. One is a recent deep reinforcement learning\nmethod based on an actor-critic algorithm, Deep Deterministic Policy Gradient\n(DDPG), that has been shown to perform well on various control benchmarks. The\nother one is a direct policy search method, Covariance Matrix Adaptation\nEvolution Strategy (CMA-ES), a black-box optimization method that is widely\nused for robot learning. The algorithms are evaluated on a continuous version\nof the mountain car benchmark problem, so as to compare their sample\ncomplexity. From a preliminary analysis, we expect DDPG to be more sample\nefficient than CMA-ES, which is confirmed by our experimental results.","url_abs":"http://arxiv.org/abs/1606.09152v2","url_pdf":"http://arxiv.org/pdf/1606.09152v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"actor-critic-versus-direct-policy-search-a","repo_url":"https://github.com/MOCR/DDPG","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"deep-reinforcement-learning","task_name":"Deep Reinforcement Learning"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"ddpg","method_name":"DDPG"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"experience-replay","method_name":"Experience Replay"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"weight-decay","method_name":"Weight Decay"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}