{"title":"元学习的难题在于学习什么。","authors":"Yosef Prat, Ehud Lamm","doi":"10.1017/S0140525X24000268","DOIUrl":null,"url":null,"abstract":"<p><p>Binz et al. highlight the potential of meta-learning to greatly enhance the flexibility of AI algorithms, as well as to approximate human behavior more accurately than traditional learning methods. We wish to emphasize a basic problem that lies underneath these two objectives, and in turn suggest another perspective of the required notion of \"meta\" in meta-learning: knowing what to learn.</p>","PeriodicalId":8698,"journal":{"name":"Behavioral and Brain Sciences","volume":null,"pages":null},"PeriodicalIF":16.6000,"publicationDate":"2024-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"The hard problem of meta-learning is what-to-learn.\",\"authors\":\"Yosef Prat, Ehud Lamm\",\"doi\":\"10.1017/S0140525X24000268\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>Binz et al. highlight the potential of meta-learning to greatly enhance the flexibility of AI algorithms, as well as to approximate human behavior more accurately than traditional learning methods. We wish to emphasize a basic problem that lies underneath these two objectives, and in turn suggest another perspective of the required notion of \\\"meta\\\" in meta-learning: knowing what to learn.</p>\",\"PeriodicalId\":8698,\"journal\":{\"name\":\"Behavioral and Brain Sciences\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":16.6000,\"publicationDate\":\"2024-09-23\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Behavioral and Brain Sciences\",\"FirstCategoryId\":\"102\",\"ListUrlMain\":\"https://doi.org/10.1017/S0140525X24000268\",\"RegionNum\":1,\"RegionCategory\":\"心理学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"BEHAVIORAL SCIENCES\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Behavioral and Brain Sciences","FirstCategoryId":"102","ListUrlMain":"https://doi.org/10.1017/S0140525X24000268","RegionNum":1,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"BEHAVIORAL SCIENCES","Score":null,"Total":0}
The hard problem of meta-learning is what-to-learn.
Binz et al. highlight the potential of meta-learning to greatly enhance the flexibility of AI algorithms, as well as to approximate human behavior more accurately than traditional learning methods. We wish to emphasize a basic problem that lies underneath these two objectives, and in turn suggest another perspective of the required notion of "meta" in meta-learning: knowing what to learn.
期刊介绍:
Behavioral and Brain Sciences (BBS) is a highly respected journal that employs an innovative approach called Open Peer Commentary. This format allows for the publication of noteworthy and contentious research from various fields including psychology, neuroscience, behavioral biology, and cognitive science. Each article is accompanied by 20-40 commentaries from experts across these disciplines, as well as a response from the author themselves. This unique setup creates a captivating forum for the exchange of ideas, critical analysis, and the integration of research within the behavioral and brain sciences, spanning topics from molecular neurobiology and artificial intelligence to the philosophy of the mind.