{"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/reinforcement-learning-via-recurrent","title":"Reinforcement Learning via Recurrent Convolutional Neural Networks","arxiv_id":"1701.02392","date":"2017-01-09","proceeding":null,"authors":["Tanmay Shankar","Santosha K. Dwivedy","Prithwijit Guha"],"abstract":"Deep Reinforcement Learning has enabled the learning of policies for complex\ntasks in partially observable environments, without explicitly learning the\nunderlying model of the tasks. While such model-free methods achieve\nconsiderable performance, they often ignore the structure of task. We present a\nnatural representation of to Reinforcement Learning (RL) problems using\nRecurrent Convolutional Neural Networks (RCNNs), to better exploit this\ninherent structure. We define 3 such RCNNs, whose forward passes execute an\nefficient Value Iteration, propagate beliefs of state in partially observable\nenvironments, and choose optimal actions respectively. Backpropagating\ngradients through these RCNNs allows the system to explicitly learn the\nTransition Model and Reward Function associated with the underlying MDP,\nserving as an elegant alternative to classical model-based RL. We evaluate the\nproposed algorithms in simulation, considering a robot planning problem. We\ndemonstrate the capability of our framework to reduce the cost of replanning,\nlearn accurate MDP models, and finally re-plan with learnt models to achieve\nnear-optimal policies.","url_abs":"http://arxiv.org/abs/1701.02392v1","url_pdf":"http://arxiv.org/pdf/1701.02392v1.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":"reinforcement-learning-via-recurrent","repo_url":"https://github.com/tanmayshankar/RCNN_MDP","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"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":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}