{"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/gdi-rethinking-what-makes-reinforcement","title":"GDI: Rethinking What Makes Reinforcement Learning Different From Supervised Learning","arxiv_id":"2106.06232","date":"2021-06-11","proceeding":null,"authors":["Jiajun Fan","Changnan Xiao","Yue Huang"],"abstract":"Deep Q Network (DQN) firstly kicked the door of deep reinforcement learning (DRL) via combining deep learning (DL) with reinforcement learning (RL), which has noticed that the distribution of the acquired data would change during the training process. DQN found this property might cause instability for training, so it proposed effective methods to handle the downside of the property. Instead of focusing on the unfavourable aspects, we find it critical for RL to ease the gap between the estimated data distribution and the ground truth data distribution while supervised learning (SL) fails to do so. From this new perspective, we extend the basic paradigm of RL called the Generalized Policy Iteration (GPI) into a more generalized version, which is called the Generalized Data Distribution Iteration (GDI). We see massive RL algorithms and techniques can be unified into the GDI paradigm, which can be considered as one of the special cases of GDI. We provide theoretical proof of why GDI is better than GPI and how it works. Several practical algorithms based on GDI have been proposed to verify the effectiveness and extensiveness of it. Empirical experiments prove our state-of-the-art (SOTA) performance on Arcade Learning Environment (ALE), wherein our algorithm has achieved 9620.98% mean human normalized score (HNS), 1146.39% median HNS and 22 human world record breakthroughs (HWRB) using only 200M training frames. Our work aims to lead the RL research to step into the journey of conquering the human world records and seek real superhuman agents on both performance and efficiency.","url_abs":"https://arxiv.org/abs/2106.06232v6","url_pdf":"https://arxiv.org/pdf/2106.06232v6.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":[],"tasks":[{"task_slug":"atari-games","task_name":"Atari Games"},{"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":[{"leaderboard":"/sota/atari-games-on-atari-2600-beam-rider","task":"Atari Games","dataset":"Atari 2600 Beam Rider","model":"GDI-I3","rank_in_archive_order":6,"of":49,"metrics":{"Score":"162100"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-berzerk","task":"Atari Games","dataset":"Atari 2600 Berzerk","model":"GDI-I3","rank_in_archive_order":9,"of":39,"metrics":{"Score":"7607"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-bowling","task":"Atari Games","dataset":"Atari 2600 Bowling","model":"GDI-I3","rank_in_archive_order":6,"of":44,"metrics":{"Score":"201.9"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-boxing","task":"Atari Games","dataset":"Atari 2600 Boxing","model":"GDI-H3","rank_in_archive_order":7,"of":45,"metrics":{"Score":"100"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-centipede","task":"Atari Games","dataset":"Atari 2600 Centipede","model":"GDI-I3","rank_in_archive_order":8,"of":45,"metrics":{"Score":"155830"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-chopper-command","task":"Atari Games","dataset":"Atari 2600 Chopper Command","model":"GDI-H3","rank_in_archive_order":1,"of":45,"metrics":{"Score":"999999"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-crazy-climber","task":"Atari Games","dataset":"Atari 2600 Crazy Climber","model":"GDI-I3","rank_in_archive_order":8,"of":49,"metrics":{"Score":"201000"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-defender","task":"Atari Games","dataset":"Atari 2600 Defender","model":"GDI-I3","rank_in_archive_order":3,"of":21,"metrics":{"Score":"893110"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-demon-attack","task":"Atari Games","dataset":"Atari 2600 Demon Attack","model":"GDI-I3","rank_in_archive_order":2,"of":46,"metrics":{"Score":"675530"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-double-dunk","task":"Atari Games","dataset":"Atari 2600 Double Dunk","model":"GDI-H3","rank_in_archive_order":2,"of":43,"metrics":{"Score":"24"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-enduro","task":"Atari Games","dataset":"Atari 2600 Enduro","model":"GDI-I3","rank_in_archive_order":1,"of":48,"metrics":{"Score":"14330"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-fishing-derby","task":"Atari Games","dataset":"Atari 2600 Fishing Derby","model":"GDI-I3","rank_in_archive_order":7,"of":44,"metrics":{"Score":"59"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-freeway","task":"Atari Games","dataset":"Atari 2600 Freeway","model":"GDI-I3","rank_in_archive_order":7,"of":59,"metrics":{"Score":"34"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-frostbite","task":"Atari Games","dataset":"Atari 2600 Frostbite","model":"GDI-I3","rank_in_archive_order":10,"of":53,"metrics":{"Score":"10485"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-gravitar","task":"Atari Games","dataset":"Atari 2600 Gravitar","model":"GDI-I3","rank_in_archive_order":8,"of":53,"metrics":{"Score":"5905"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-hero","task":"Atari Games","dataset":"Atari 2600 HERO","model":"GDI-I3","rank_in_archive_order":5,"of":44,"metrics":{"Score":"38330"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-ice-hockey","task":"Atari Games","dataset":"Atari 2600 Ice Hockey","model":"GDI-I3","rank_in_archive_order":5,"of":43,"metrics":{"Score":"44.94"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-james-bond","task":"Atari Games","dataset":"Atari 2600 James Bond","model":"GDI-I3","rank_in_archive_order":2,"of":45,"metrics":{"Score":"594500"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-kangaroo","task":"Atari Games","dataset":"Atari 2600 Kangaroo","model":"GDI-I3","rank_in_archive_order":10,"of":47,"metrics":{"Score":"14500"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-krull","task":"Atari Games","dataset":"Atari 2600 Krull","model":"GDI-I3","rank_in_archive_order":5,"of":45,"metrics":{"Score":"97575"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-montezumas-revenge","task":"Atari Games","dataset":"Atari 2600 Montezuma's Revenge","model":"GDI-I3","rank_in_archive_order":11,"of":50,"metrics":{"Score":"3000"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-ms-pacman","task":"Atari Games","dataset":"Atari 2600 Ms. Pacman","model":"GDI-I3","rank_in_archive_order":7,"of":47,"metrics":{"Score":"11536"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-name-this-game","task":"Atari Games","dataset":"Atari 2600 Name This Game","model":"GDI-I3","rank_in_archive_order":6,"of":43,"metrics":{"Score":"34434"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-phoenix","task":"Atari Games","dataset":"Atari 2600 Phoenix","model":"GDI-I3","rank_in_archive_order":4,"of":21,"metrics":{"Score":"894460"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-pitfall","task":"Atari Games","dataset":"Atari 2600 Pitfall!","model":"GDI-I3","rank_in_archive_order":14,"of":23,"metrics":{"Score":"0"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-private-eye","task":"Atari Games","dataset":"Atari 2600 Private Eye","model":"GDI-I3","rank_in_archive_order":5,"of":52,"metrics":{"Score":"15100"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-qbert","task":"Atari Games","dataset":"Atari 2600 Q*Bert","model":"GDI-I3","rank_in_archive_order":14,"of":57,"metrics":{"Score":"27800"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-road-runner","task":"Atari Games","dataset":"Atari 2600 Road Runner","model":"GDI-I3","rank_in_archive_order":2,"of":44,"metrics":{"Score":"878600"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-robotank","task":"Atari Games","dataset":"Atari 2600 Robotank","model":"GDI-I3","rank_in_archive_order":4,"of":42,"metrics":{"Score":"108.2"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-seaquest","task":"Atari Games","dataset":"Atari 2600 Seaquest","model":"GDI-I3","rank_in_archive_order":8,"of":57,"metrics":{"Score":"943910"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-skiing","task":"Atari Games","dataset":"Atari 2600 Skiing","model":"GDI-I3","rank_in_archive_order":19,"of":23,"metrics":{"Score":"-6774"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-solaris","task":"Atari Games","dataset":"Atari 2600 Solaris","model":"GDI-I3","rank_in_archive_order":6,"of":23,"metrics":{"Score":"11074"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-space-invaders","task":"Atari Games","dataset":"Atari 2600 Space Invaders","model":"GDI-I3","rank_in_archive_order":4,"of":55,"metrics":{"Score":"140460"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-star-gunner","task":"Atari Games","dataset":"Atari 2600 Star Gunner","model":"GDI-I3","rank_in_archive_order":5,"of":43,"metrics":{"Score":"465750"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-surround","task":"Atari Games","dataset":"Atari 2600 Surround","model":"GDI-I3","rank_in_archive_order":14,"of":15,"metrics":{"Score":"-7.8"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-tennis","task":"Atari Games","dataset":"Atari 2600 Tennis","model":"GDI-I3","rank_in_archive_order":1,"of":43,"metrics":{"Score":"24"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-time-pilot","task":"Atari Games","dataset":"Atari 2600 Time Pilot","model":"GDI-I3","rank_in_archive_order":6,"of":44,"metrics":{"Score":"216770"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-tutankham","task":"Atari Games","dataset":"Atari 2600 Tutankham","model":"GDI-I3","rank_in_archive_order":3,"of":44,"metrics":{"Score":"423.9"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-2600-up-and-down","task":"Atari Games","dataset":"Atari 2600 Up and Down","model":"GDI-I3","rank_in_archive_order":1,"of":44,"metrics":{"Score":"986440"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-57","task":"Atari Games","dataset":"Atari-57","model":"GDI-H3(200M frames)","rank_in_archive_order":2,"of":11,"metrics":{"Human World Record Breakthrough":"22","Mean Human Normalized Score":"9620.98%"},"uses_additional_data":false},{"leaderboard":"/sota/atari-games-on-atari-57","task":"Atari Games","dataset":"Atari-57","model":"GDI-H3","rank_in_archive_order":11,"of":11,"metrics":{},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}