{"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/learning-model-based-planning-from-scratch","title":"Learning model-based planning from scratch","arxiv_id":"1707.06170","date":"2017-07-19","proceeding":null,"authors":["Razvan Pascanu","Yujia Li","Oriol Vinyals","Nicolas Heess","Lars Buesing","Sebastien Racanière","David Reichert","Théophane Weber","Daan Wierstra","Peter Battaglia"],"abstract":"Conventional wisdom holds that model-based planning is a powerful approach to\nsequential decision-making. It is often very challenging in practice, however,\nbecause while a model can be used to evaluate a plan, it does not prescribe how\nto construct a plan. Here we introduce the \"Imagination-based Planner\", the\nfirst model-based, sequential decision-making agent that can learn to\nconstruct, evaluate, and execute plans. Before any action, it can perform a\nvariable number of imagination steps, which involve proposing an imagined\naction and evaluating it with its model-based imagination. All imagined actions\nand outcomes are aggregated, iteratively, into a \"plan context\" which\nconditions future real and imagined actions. The agent can even decide how to\nimagine: testing out alternative imagined actions, chaining sequences of\nactions together, or building a more complex \"imagination tree\" by navigating\nflexibly among the previously imagined states using a learned policy. And our\nagent can learn to plan economically, jointly optimizing for external rewards\nand computational costs associated with using its imagination. We show that our\narchitecture can learn to solve a challenging continuous control problem, and\nalso learn elaborate planning strategies in a discrete maze-solving task. Our\nwork opens a new direction toward learning the components of a model-based\nplanning system and how to use them.","url_abs":"http://arxiv.org/abs/1707.06170v1","url_pdf":"http://arxiv.org/pdf/1707.06170v1.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":"learning-model-based-planning-from-scratch","repo_url":"https://github.com/Driesssens/ibp-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"learning-model-based-planning-from-scratch","repo_url":"https://github.com/obster991/RL-model-based-planning-from-scratch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"continuous-control","task_name":"Continuous Control"},{"task_slug":"decision-making","task_name":"Decision Making"},{"task_slug":"sequential-decision-making","task_name":"Sequential Decision Making"},{"task_slug":"continuous-control","task_name":"continuous-control"},{"task_slug":"model","task_name":"model"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1707.06170","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}