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

相关概念视频

Language and Cognition01:27

Language and Cognition

340
Language serves as a bridge between ideas and communication, influencing how individuals perceive and interact with the world. Psychologists have long debated whether language shapes thought or vice versa. This discussion gained grip with Edward Sapir and Benjamin Lee Whorf in the 1940s, who proposed that language determines thought, a concept known as linguistic determinism. They suggested that the vocabulary and structure of a language influence how its speakers think and perceive reality.
340

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

The human language processing system straightens natural speech.

bioRxiv : the preprint server for biology·2026
Same author

Does the brain really know what word is coming next?

eLife·2026
Same author

Decoding concept representations in aphasia after stroke.

bioRxiv : the preprint server for biology·2026
Same author

Rapid cortical mapping with cross-participant encoding models.

bioRxiv : the preprint server for biology·2026
Same author

Visual Semantic Encoding and Identification of Naturalistic Movies via High-Density Diffuse Optical Tomography.

bioRxiv : the preprint server for biology·2025
Same author

Evaluating scientific theories as predictive models in language neuroscience.

bioRxiv : the preprint server for biology·2025

相关实验视频

Updated: Jun 20, 2025

Measuring Statistical Learning Across Modalities and Domains in School-Aged Children Via an Online Platform and Neuroimaging Techniques
08:05

Measuring Statistical Learning Across Modalities and Domains in School-Aged Children Via an Online Platform and Neuroimaging Techniques

Published on: June 30, 2020

7.5K

在fMRI中的语言编码模型的缩放规律.

Richard J Antonello1, Aditya R Vaidya1, Alexander G Huth2

  • 1Department of Computer Science, The University of Texas at Austin.

Advances in neural information processing systems
|July 22, 2024
PubMed
概括

较大的语言模型显著改善了在自然语言处理过程中大脑活动的预测. 性能与模型大小对数推移,在某些大脑区域接近理论极限.

科学领域:

  • 神经科学是一个神经科学.
  • 计算语言学 计算语言学
  • 人工智能的人工智能

背景情况:

  • 基于变压器的语言模型擅长预测大脑对语言的反应.
  • 之前的研究主要使用较小的模型,如GPT-2.
  • 更大的开源模型的潜力在很大程度上仍未被探索.

研究的目的:

  • 调查较大的开源语言模型 (OPT,LLaMA) 与较小的模型相比,是否可以改善大脑反应预测.
  • 分析模型大小和预测性能的训练数据的缩放特性.
  • 评估声学编码模型 (HuBERT,WavLM,Whisper) 的性能及其扩展行为.

主要方法:

  • fMRI数据被用来记录大脑对自然语言的反应.
  • 使用了大型开源语言模型 (OPT,LLaMA) 和声学模型 (HuBERT,WavLM,Whisper).
  • 编码模型经过训练,并对其预测大脑活动的能力进行评估.
  • 通过与持有测试数据的相关性来测量性能,并分析模型和数据大小的缩放.

主要成果:

  • 大脑预测性能显示了对模型大小的对数缩放,从125M到30B的参数.
  • 在3个受试者中观察到编码性能增加了~15%.

更多相关视频

fMRI Mapping of Brain Activity Associated with the Vocal Production of Consonant and Dissonant Intervals
11:15

fMRI Mapping of Brain Activity Associated with the Vocal Production of Consonant and Dissonant Intervals

Published on: May 23, 2017

7.2K
Functional Magnetic Resonance Imaging fMRI with Auditory Stimulation in Songbirds
13:05

Functional Magnetic Resonance Imaging fMRI with Auditory Stimulation in Songbirds

Published on: June 3, 2013

18.2K

相关实验视频

Last Updated: Jun 20, 2025

Measuring Statistical Learning Across Modalities and Domains in School-Aged Children Via an Online Platform and Neuroimaging Techniques
08:05

Measuring Statistical Learning Across Modalities and Domains in School-Aged Children Via an Online Platform and Neuroimaging Techniques

Published on: June 30, 2020

7.5K
fMRI Mapping of Brain Activity Associated with the Vocal Production of Consonant and Dissonant Intervals
11:15

fMRI Mapping of Brain Activity Associated with the Vocal Production of Consonant and Dissonant Intervals

Published on: May 23, 2017

7.2K
Functional Magnetic Resonance Imaging fMRI with Auditory Stimulation in Songbirds
13:05

Functional Magnetic Resonance Imaging fMRI with Auditory Stimulation in Songbirds

Published on: June 3, 2013

18.2K
  • 当增加fMRI训练数据集的大小时,发现了类似的对数缩放.
  • 声学编码模型显示,随着模型大小的增加,同样可以比较的改进.
  • 对噪声上限的分析表明,在前皮质和听觉皮质等区域,性能正在接近理论最大值.

    • 结论:

    • 增加语言模型和训练数据的规模显著提高了对语言大脑反应的预测.
    • 这些发现表明,大规模模型正在接近语言处理当前神经成像数据的极限.
    • 这项研究为改善对大脑语言处理和先进解码应用的科学理解铺平了道路.