{"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/the-fast-and-the-flexible-training-neural","title":"The Fast and the Flexible: training neural networks to learn to follow instructions from small data","arxiv_id":"1809.06194","date":"2018-09-17","proceeding":"WS 2019 5","authors":["Rezka Leonandya","Elia Bruni","Dieuwke Hupkes","Germán Kruszewski"],"abstract":"Learning to follow human instructions is a long-pursued goal in artificial\nintelligence. The task becomes particularly challenging if no prior knowledge\nof the employed language is assumed while relying only on a handful of examples\nto learn from. Work in the past has relied on hand-coded components or manually\nengineered features to provide strong inductive biases that make learning in\nsuch situations possible. In contrast, here we seek to establish whether this\nknowledge can be acquired automatically by a neural network system through a\ntwo phase training procedure: A (slow) offline learning stage where the network\nlearns about the general structure of the task and a (fast) online adaptation\nphase where the network learns the language of a new given speaker. Controlled\nexperiments show that when the network is exposed to familiar instructions but\ncontaining novel words, the model adapts very efficiently to the new\nvocabulary. Moreover, even for human speakers whose language usage can depart\nsignificantly from our artificial training language, our network can still make\nuse of its automatically acquired inductive bias to learn to follow\ninstructions more effectively.","url_abs":"http://arxiv.org/abs/1809.06194v2","url_pdf":"http://arxiv.org/pdf/1809.06194v2.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":"the-fast-and-the-flexible-training-neural","repo_url":"https://github.com/rezkaaufar/fast-and-flexible","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"inductive-bias","task_name":"Inductive Bias"},{"task_slug":"small-data","task_name":"Small Data Image Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1809.06194","atlas_url":"https://app.syntology.ai/?focus=1809.06194","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}