Jove
Visualize
联系我们
JoVE
x logofacebook logolinkedin logoyoutube logo
关于 JoVE
概览领导团队博客JoVE 帮助中心
作者
出版流程编辑委员会范围与政策同行评审常见问题投稿
图书馆员
用户评价订阅访问资源图书馆顾问委员会常见问题
研究
JoVE JournalMethods CollectionsJoVE Encyclopedia of Experiments存档
教育
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab Manual教师资源中心教师网站
使用条款与条件
隐私政策
政策

相关实验视频

Updated: Mar 3, 2026

Fabrication of Three-Dimensional Graphene-Based Polyhedrons via Origami-Like Self-Folding
14:52

Fabrication of Three-Dimensional Graphene-Based Polyhedrons via Origami-Like Self-Folding

Published on: September 23, 2018

9.4K

材料科学中的大型语言模型:通过石墨烯合成来评估RAG评估框架.

Zen Han Cho1, Matthew Osvaldo1, Sayan Doloi1

  • 1School of Materials Science and Engineering, Nanyang Technological University 50 Nanyang Avenue 639798 Singapore leonard.ngwt@ntu.edu.sg.

RSC advances
|March 2, 2026
PubMed
概括

相关概念视频

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序
Same author

A green and sustainable approach using <i>Thuja occidentalis</i>-mediated Fe<sub>3</sub>O<sub>4</sub> QDs decorated on rGO NSs for enhanced photocatalytic degradation of antibiotics.

RSC advances·2026
Same author

Deployment of a cloud-based passive defecation monitoring system for continuous gut health monitoring.

Nature protocols·2026
Same author

Self-Driving Laboratories: Translating Materials Science from Laboratory to Factory.

ACS omega·2025
Same author

Lightweight error-tolerant edge detection using memristor-enabled stochastic computing.

Nature communications·2025
Same author

Cardiovascular outcomes and safety of SGLT2 inhibitors in chronic kidney disease patients.

Frontiers in endocrinology·2023
Same author

Symbionts in <i>Hodgkinia</i>-free cicadas and their implications for co-evolution between endosymbionts and host insects.

Applied and environmental microbiology·2023

在科学中评估回收增强生成 (RAG) 系统是很困难的. 与BERTScore或LLM-as-a-Judge不同,RAGAS显示出对评估科学RAG绩效的希望,但仍然存在局限性.

科学领域:

  • 材料科学 材料科学 材料科学
  • 石墨烯合成的合成方法

背景情况:

  • 由于技术复杂性和精度需求,在专业的科学领域评估检索增强生成 (RAG) 系统具有挑战性.
  • 现有的自动化评估框架可能无法充分捕捉RAG在科学环境中的表现.

研究的目的:

  • 系统地分析和比较科学RAG系统的自动评估框架.
  • 建立用于在专业研究领域评估人工智能的方法指南.

主要方法:

  • 使用石墨烯合成作为案例研究,制定了一个全面的评估协议.
  • 他们比较了四种评估方法:RAGAS,BERTScore,LLM-as-a-Judge和专家人类评估.
  • 通过20个特定领域的问题来评估绩效.

主要成果:

  • 在科学RAG评估中,BERTScore缺乏可解释性和评分敏感性.
  • 作为法官的法学士没有捕捉到提取增强的好处.
  • RAGAS成功地确定了从检索增强中获得的相对性能改善,特别是在较小的模型中,但在科学内容的绝对得分解释方面存在局限性.

结论:

  • RAGAS展示了评估科学RAG系统的潜力,提供了对检索增强效益的见解.

更多相关视频

Synthesis and Functionalization of 3D Nano-graphene Materials: Graphene Aerogels and Graphene Macro Assemblies
10:23

Synthesis and Functionalization of 3D Nano-graphene Materials: Graphene Aerogels and Graphene Macro Assemblies

Published on: November 5, 2015

14.6K
Advanced Self-Healing Asphalt Reinforced by Graphene Structures: An Atomistic Insight
08:03

Advanced Self-Healing Asphalt Reinforced by Graphene Structures: An Atomistic Insight

Published on: May 31, 2022

5.7K

相关实验视频

Last Updated: Mar 3, 2026

Fabrication of Three-Dimensional Graphene-Based Polyhedrons via Origami-Like Self-Folding
14:52

Fabrication of Three-Dimensional Graphene-Based Polyhedrons via Origami-Like Self-Folding

Published on: September 23, 2018

9.4K
Synthesis and Functionalization of 3D Nano-graphene Materials: Graphene Aerogels and Graphene Macro Assemblies
10:23

Synthesis and Functionalization of 3D Nano-graphene Materials: Graphene Aerogels and Graphene Macro Assemblies

Published on: November 5, 2015

14.6K
Advanced Self-Healing Asphalt Reinforced by Graphene Structures: An Atomistic Insight
08:03

Advanced Self-Healing Asphalt Reinforced by Graphene Structures: An Atomistic Insight

Published on: May 31, 2022

5.7K
  • 自动化框架需要进一步开发,以准确评估AI在专业科学领域的性能.
  • 这些发现为在科学中部署人工智能的研究人员提供了关键考虑.