{"title":"The Importance of Understanding Deep Learning.","authors":"Tim Räz, Claus Beisbart","doi":"10.1007/s10670-022-00605-y","DOIUrl":null,"url":null,"abstract":"<p><p>Some machine learning models, in particular deep neural networks (DNNs), are not very well understood; nevertheless, they are frequently used in science. Does this lack of understanding pose a problem for using DNNs to understand empirical phenomena? Emily Sullivan has recently argued that understanding with DNNs is not limited by our lack of understanding of DNNs themselves. In the present paper, we will argue, <i>contra</i> Sullivan, that our current lack of understanding of DNNs does limit our ability to understand with DNNs. Sullivan's claim hinges on which notion of understanding is at play. If we employ a weak notion of understanding, then her claim is tenable, but rather weak. If, however, we employ a strong notion of understanding, particularly explanatory understanding, then her claim is not tenable.</p>","PeriodicalId":10165,"journal":{"name":"Climate Dynamics","volume":"26 1","pages":"1823-1840"},"PeriodicalIF":3.8000,"publicationDate":"2024-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11090801/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Climate Dynamics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1007/s10670-022-00605-y","RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2022/8/7 0:00:00","PubModel":"Epub","JCR":"Q2","JCRName":"METEOROLOGY & ATMOSPHERIC SCIENCES","Score":null,"Total":0}
引用次数: 0
Abstract
Some machine learning models, in particular deep neural networks (DNNs), are not very well understood; nevertheless, they are frequently used in science. Does this lack of understanding pose a problem for using DNNs to understand empirical phenomena? Emily Sullivan has recently argued that understanding with DNNs is not limited by our lack of understanding of DNNs themselves. In the present paper, we will argue, contra Sullivan, that our current lack of understanding of DNNs does limit our ability to understand with DNNs. Sullivan's claim hinges on which notion of understanding is at play. If we employ a weak notion of understanding, then her claim is tenable, but rather weak. If, however, we employ a strong notion of understanding, particularly explanatory understanding, then her claim is not tenable.
期刊介绍:
The international journal Climate Dynamics provides for the publication of high-quality research on all aspects of the dynamics of the global climate system.
Coverage includes original paleoclimatic, diagnostic, analytical and numerical modeling research on the structure and behavior of the atmosphere, oceans, cryosphere, biomass and land surface as interacting components of the dynamics of global climate. Contributions are focused on selected aspects of climate dynamics on particular scales of space or time.
The journal also publishes reviews and papers emphasizing an integrated view of the physical and biogeochemical processes governing climate and climate change.