人类的组合方式是元学习的吗？

IF 19.3 1区材料科学 Q1 CHEMISTRY, PHYSICAL

ACS Energy Letters Pub Date : 2024-09-23 DOI:10.1017/S0140525X24000189

Jacob Russin, Sam Whitman McGrath, Ellie Pavlick, Michael J Frank

{"title":"人类的组合方式是元学习的吗？","authors":"Jacob Russin, Sam Whitman McGrath, Ellie Pavlick, Michael J Frank","doi":"10.1017/S0140525X24000189","DOIUrl":null,"url":null,"abstract":"Recent studies suggest that meta-learning may provide an original solution to an enduring puzzle about whether neural networks can explain compositionality - in particular, by raising the prospect that compositionality can be understood as an emergent property of an inner-loop learning algorithm. We elaborate on this hypothesis and consider its empirical predictions regarding the neural mechanisms and development of human compositionality.","PeriodicalId":16,"journal":{"name":"ACS Energy Letters ","volume":null,"pages":null},"PeriodicalIF":19.3000,"publicationDate":"2024-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Is human compositionality meta-learned?\",\"authors\":\"Jacob Russin, Sam Whitman McGrath, Ellie Pavlick, Michael J Frank\",\"doi\":\"10.1017/S0140525X24000189\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Recent studies suggest that meta-learning may provide an original solution to an enduring puzzle about whether neural networks can explain compositionality - in particular, by raising the prospect that compositionality can be understood as an emergent property of an inner-loop learning algorithm. We elaborate on this hypothesis and consider its empirical predictions regarding the neural mechanisms and development of human compositionality.\",\"PeriodicalId\":16,\"journal\":{\"name\":\"ACS Energy Letters \",\"volume\":null,\"pages\":null},\"PeriodicalIF\":19.3000,\"publicationDate\":\"2024-09-23\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"ACS Energy Letters \",\"FirstCategoryId\":\"102\",\"ListUrlMain\":\"https://doi.org/10.1017/S0140525X24000189\",\"RegionNum\":1,\"RegionCategory\":\"材料科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"CHEMISTRY, PHYSICAL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"ACS Energy Letters ","FirstCategoryId":"102","ListUrlMain":"https://doi.org/10.1017/S0140525X24000189","RegionNum":1,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"CHEMISTRY, PHYSICAL","Score":null,"Total":0}

引用次数: 0

摘要

最近的研究表明，元学习（meta-learning）可能会为神经网络能否解释合成性这一长期谜题提供一种新的解决方案，特别是通过提出合成性可以被理解为内环学习算法的一种新兴属性这一前景。我们将详细阐述这一假设，并考虑它对人类构图性的神经机制和发展的经验预测。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

Is human compositionality meta-learned?

Recent studies suggest that meta-learning may provide an original solution to an enduring puzzle about whether neural networks can explain compositionality - in particular, by raising the prospect that compositionality can be understood as an emergent property of an inner-loop learning algorithm. We elaborate on this hypothesis and consider its empirical predictions regarding the neural mechanisms and development of human compositionality.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

ACS Energy Letters Energy-Renewable Energy, Sustainability and the Environment

CiteScore

31.20

自引率

5.00%

发文量

469

审稿时长

1 months

期刊介绍： ACS Energy Letters is a monthly journal that publishes papers reporting new scientific advances in energy research. The journal focuses on topics that are of interest to scientists working in the fundamental and applied sciences. Rapid publication is a central criterion for acceptance, and the journal is known for its quick publication times, with an average of 4-6 weeks from submission to web publication in As Soon As Publishable format. ACS Energy Letters is ranked as the number one journal in the Web of Science Electrochemistry category. It also ranks within the top 10 journals for Physical Chemistry, Energy & Fuels, and Nanoscience & Nanotechnology. The journal offers several types of articles, including Letters, Energy Express, Perspectives, Reviews, Editorials, Viewpoints and Energy Focus. Additionally, authors have the option to submit videos that summarize or support the information presented in a Perspective or Review article, which can be highlighted on the journal's website. ACS Energy Letters is abstracted and indexed in Chemical Abstracts Service/SciFinder, EBSCO-summon, PubMed, Web of Science, Scopus and Portico.