{"title":"How Well Do Large Language Models Understand Tables in Materials Science?","authors":"Defne Circi, Ghazal Khalighinejad, Anlan Chen, Bhuwan Dhingra, L. Catherine Brinson","doi":"10.1007/s40192-024-00362-6","DOIUrl":null,"url":null,"abstract":"<p>Advances in materials science require leveraging past findings and data from the vast published literature. While some materials data repositories are being built, they typically rely on newly created data in narrow domains because extracting detailed data and metadata from the enormous wealth of publications is immensely challenging. The advent of large language models (LLMs) presents a new opportunity to rapidly and accurately extract data and insights from the published literature and transform it into structured data formats for easy query and reuse. In this paper, we build on initial strategies for using LLMs for rapid and autonomous data extraction from materials science articles in a format curatable by materials databases. We presented the subdomain of polymer composites as our example use case and demonstrated the success and challenges of LLMs on extracting tabular data. We explored different table representations for use with LLMs, finding that a multimodal model with an image input yielded the most promising results. This model achieved an accuracy score of 0.910 for composition information extraction and an F<span>\\(_1\\)</span> score of 0.863 for property name information extraction. With the most conservative evaluation for the property extraction requiring exact match in all the details, we obtained an F<span>\\(_1\\)</span> score of 0.419. We observed that by allowing varying degrees of flexibility in the evaluation, the score can increase to 0.769. We envision that the results and analysis from this study will promote further research directions in developing information extraction strategies from materials information sources.</p>","PeriodicalId":13604,"journal":{"name":"Integrating Materials and Manufacturing Innovation","volume":"1 1","pages":""},"PeriodicalIF":2.4000,"publicationDate":"2024-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Integrating Materials and Manufacturing Innovation","FirstCategoryId":"88","ListUrlMain":"https://doi.org/10.1007/s40192-024-00362-6","RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ENGINEERING, MANUFACTURING","Score":null,"Total":0}
引用次数: 0
Abstract
Advances in materials science require leveraging past findings and data from the vast published literature. While some materials data repositories are being built, they typically rely on newly created data in narrow domains because extracting detailed data and metadata from the enormous wealth of publications is immensely challenging. The advent of large language models (LLMs) presents a new opportunity to rapidly and accurately extract data and insights from the published literature and transform it into structured data formats for easy query and reuse. In this paper, we build on initial strategies for using LLMs for rapid and autonomous data extraction from materials science articles in a format curatable by materials databases. We presented the subdomain of polymer composites as our example use case and demonstrated the success and challenges of LLMs on extracting tabular data. We explored different table representations for use with LLMs, finding that a multimodal model with an image input yielded the most promising results. This model achieved an accuracy score of 0.910 for composition information extraction and an F\(_1\) score of 0.863 for property name information extraction. With the most conservative evaluation for the property extraction requiring exact match in all the details, we obtained an F\(_1\) score of 0.419. We observed that by allowing varying degrees of flexibility in the evaluation, the score can increase to 0.769. We envision that the results and analysis from this study will promote further research directions in developing information extraction strategies from materials information sources.
期刊介绍:
The journal will publish: Research that supports building a model-based definition of materials and processes that is compatible with model-based engineering design processes and multidisciplinary design optimization; Descriptions of novel experimental or computational tools or data analysis techniques, and their application, that are to be used for ICME; Best practices in verification and validation of computational tools, sensitivity analysis, uncertainty quantification, and data management, as well as standards and protocols for software integration and exchange of data; In-depth descriptions of data, databases, and database tools; Detailed case studies on efforts, and their impact, that integrate experiment and computation to solve an enduring engineering problem in materials and manufacturing.