{"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/policy-distillation","title":"Policy Distillation","arxiv_id":"1511.06295","date":"2015-11-19","proceeding":null,"authors":["Andrei A. Rusu","Sergio Gomez Colmenarejo","Caglar Gulcehre","Guillaume Desjardins","James Kirkpatrick","Razvan Pascanu","Volodymyr Mnih","Koray Kavukcuoglu","Raia Hadsell"],"abstract":"Policies for complex visual tasks have been successfully learned with deep\nreinforcement learning, using an approach called deep Q-networks (DQN), but\nrelatively large (task-specific) networks and extensive training are needed to\nachieve good performance. In this work, we present a novel method called policy\ndistillation that can be used to extract the policy of a reinforcement learning\nagent and train a new network that performs at the expert level while being\ndramatically smaller and more efficient. Furthermore, the same method can be\nused to consolidate multiple task-specific policies into a single policy. We\ndemonstrate these claims using the Atari domain and show that the multi-task\ndistilled agent outperforms the single-task teachers as well as a\njointly-trained DQN agent.","url_abs":"http://arxiv.org/abs/1511.06295v2","url_pdf":"http://arxiv.org/pdf/1511.06295v2.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":"policy-distillation","repo_url":"https://github.com/dsapandora/s_cera","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[{"task_slug":"deep-reinforcement-learning","task_name":"Deep Reinforcement Learning"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"},{"task_slug":"reinforcement-learning-1","task_name":"Reinforcement Learning (RL)"},{"task_slug":"reinforcement-learning-2","task_name":"reinforcement-learning"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dqn","method_name":"DQN"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"q-learning","method_name":"Q-Learning"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1511.06295","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}