{"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/been-there-done-that-meta-learning-with","title":"Been There, Done That: Meta-Learning with Episodic Recall","arxiv_id":"1805.09692","date":"2018-05-24","proceeding":"ICML 2018 7","authors":["Samuel Ritter","Jane. X. Wang","Zeb Kurth-Nelson","Siddhant M. Jayakumar","Charles Blundell","Razvan Pascanu","Matthew Botvinick"],"abstract":"Meta-learning agents excel at rapidly learning new tasks from open-ended task\ndistributions; yet, they forget what they learn about each task as soon as the\nnext begins. When tasks reoccur - as they do in natural environments -\nmetalearning agents must explore again instead of immediately exploiting\npreviously discovered solutions. We propose a formalism for generating\nopen-ended yet repetitious environments, then develop a meta-learning\narchitecture for solving these environments. This architecture melds the\nstandard LSTM working memory with a differentiable neural episodic memory. We\nexplore the capabilities of agents with this episodic LSTM in five\nmeta-learning environments with reoccurring tasks, ranging from bandits to\nnavigation and stochastic sequential decision problems.","url_abs":"http://arxiv.org/abs/1805.09692v2","url_pdf":"http://arxiv.org/pdf/1805.09692v2.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":"been-there-done-that-meta-learning-with","repo_url":"https://github.com/qihongl/dlstm-demo","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"meta-learning","task_name":"Meta-Learning"}],"methods":[{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1805.09692","atlas_url":"https://app.syntology.ai/?focus=1805.09692","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}