{"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/meta-sgd-learning-to-learn-quickly-for-few","title":"Meta-SGD: Learning to Learn Quickly for Few-Shot Learning","arxiv_id":"1707.09835","date":"2017-07-31","proceeding":null,"authors":["Zhenguo Li","Fengwei Zhou","Fei Chen","Hang Li"],"abstract":"Few-shot learning is challenging for learning algorithms that learn each task\nin isolation and from scratch. In contrast, meta-learning learns from many\nrelated tasks a meta-learner that can learn a new task more accurately and\nfaster with fewer examples, where the choice of meta-learners is crucial. In\nthis paper, we develop Meta-SGD, an SGD-like, easily trainable meta-learner\nthat can initialize and adapt any differentiable learner in just one step, on\nboth supervised learning and reinforcement learning. Compared to the popular\nmeta-learner LSTM, Meta-SGD is conceptually simpler, easier to implement, and\ncan be learned more efficiently. Compared to the latest meta-learner MAML,\nMeta-SGD has a much higher capacity by learning to learn not just the learner\ninitialization, but also the learner update direction and learning rate, all in\na single meta-learning process. Meta-SGD shows highly competitive performance\nfor few-shot learning on regression, classification, and reinforcement\nlearning.","url_abs":"http://arxiv.org/abs/1707.09835v2","url_pdf":"http://arxiv.org/pdf/1707.09835v2.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":"meta-sgd-learning-to-learn-quickly-for-few","repo_url":"https://github.com/BBDrive/Meta-SGD-RL","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"meta-sgd-learning-to-learn-quickly-for-few","repo_url":"https://github.com/ash3n/Meta-Gradients","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"meta-sgd-learning-to-learn-quickly-for-few","repo_url":"https://github.com/ash3n/Meta-SGD","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"meta-sgd-learning-to-learn-quickly-for-few","repo_url":"https://github.com/ash3n/Meta-SGD-TF","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"meta-sgd-learning-to-learn-quickly-for-few","repo_url":"https://github.com/foolyc/Meta-SGD","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"meta-sgd-learning-to-learn-quickly-for-few","repo_url":"https://github.com/llan-ml/tesp","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"meta-sgd-learning-to-learn-quickly-for-few","repo_url":"https://github.com/myungsub/meta-interpolation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"meta-sgd-learning-to-learn-quickly-for-few","repo_url":"https://github.com/tobiasvanderwerff/MetaHTR","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"meta-sgd-learning-to-learn-quickly-for-few","repo_url":"https://github.com/learnables/learn2learn","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"few-shot-learning","task_name":"Few-Shot Learning"},{"task_slug":"meta-learning","task_name":"Meta-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":"lstm","method_name":"LSTM"},{"method_slug":"maml","method_name":"MAML"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/few-shot-image-classification-on-mini-7","task":"Few-Shot Image Classification","dataset":"Mini-Imagenet 20-way (1-shot)","model":"Meta SGD","rank_in_archive_order":3,"of":6,"metrics":{"Accuracy":"17.56"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-image-classification-on-mini-7","task":"Few-Shot Image Classification","dataset":"Mini-Imagenet 20-way (1-shot)","model":"Matching Nets, (from )","rank_in_archive_order":4,"of":6,"metrics":{"Accuracy":"17.31"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-image-classification-on-mini-7","task":"Few-Shot Image Classification","dataset":"Mini-Imagenet 20-way (1-shot)","model":"Meta LSTM, (from )","rank_in_archive_order":5,"of":6,"metrics":{"Accuracy":"16.70"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-image-classification-on-mini-7","task":"Few-Shot Image Classification","dataset":"Mini-Imagenet 20-way (1-shot)","model":"MAML, (from )","rank_in_archive_order":6,"of":6,"metrics":{"Accuracy":"16.49"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-image-classification-on-mini-8","task":"Few-Shot Image Classification","dataset":"Mini-Imagenet 20-way (5-shot)","model":"Meta SGD","rank_in_archive_order":3,"of":6,"metrics":{"Accuracy":"28.92"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-image-classification-on-mini-8","task":"Few-Shot Image Classification","dataset":"Mini-Imagenet 20-way (5-shot)","model":"Meta LSTM, (from )","rank_in_archive_order":4,"of":6,"metrics":{"Accuracy":"26.06"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-image-classification-on-mini-8","task":"Few-Shot Image Classification","dataset":"Mini-Imagenet 20-way (5-shot)","model":"Matching Nets, (from )","rank_in_archive_order":5,"of":6,"metrics":{"Accuracy":"22.69"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-image-classification-on-mini-8","task":"Few-Shot Image Classification","dataset":"Mini-Imagenet 20-way (5-shot)","model":"MAML, (from )","rank_in_archive_order":6,"of":6,"metrics":{"Accuracy":"19.29"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1707.09835","atlas_url":"https://app.syntology.ai/?focus=1707.09835","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}