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

Improving Translational Accuracy02:07

Improving Translational Accuracy

2.6K
2.6K
Initiation of Translation02:33

Initiation of Translation

6.4K
6.4K
Termination of Translation01:44

Termination of Translation

5.4K
5.4K
Alternative RNA Splicing02:18

Alternative RNA Splicing

3.8K
3.8K
Lagging Strand Synthesis01:59

Lagging Strand Synthesis

13.6K
13.6K
Pre-mRNA Processing: RNA Splicing01:36

Pre-mRNA Processing: RNA Splicing

5.2K
5.2K

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

Cryo-deposited Au shells amplify nanozyme catalysis: A K<sub>m</sub>-enhanced immunoassay for mycotoxin screening.

Biosensors & bioelectronics·2026
Same author

Unlocking the prognostic power of pathomics in bladder cancer: a machine learning odyssey across multiple centers.

BMC medical imaging·2026
Same author

Zwitterionic Nanoplatform-Mediated Photopyroptosis for Abscopal Effect-Enhanced Cancer Immunotherapy.

Small (Weinheim an der Bergstrasse, Germany)·2026
Same author

Advances in 3D-bioprinted implantable liver microtissue for liver failure treatment.

Chemical communications (Cambridge, England)·2026
Same author

Engineered zwitterion-nanodelivery for precision targeting of brain metastases.

Nature communications·2026
Same author

Dual-quenched and redox-responsive gold nanoplatforms for tumor-specific multimodal theranostics.

Nanoscale·2026

相关实验视频

Updated: Jul 4, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

572

一种双阶段微调方法,用于低资源的跨语言总结.

Kaixiong Zhang1,2, Yongbing Zhang1,2, Zhengtao Yu1,2

  • 1Faculty of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650500, China.

Mathematical biosciences and engineering : MBE
|February 2, 2024
PubMed
概括

本研究引入了一种新的两阶段微调方法 (TFLCLS),通过增强多语言模型的语义理解和信息压缩来改善低资源的跨语言总结.

关键词:
跨语言的跨语言.精细调整 精细调整这是一个低资源的低资源.总结 总结 总结 总结

更多相关视频

Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody
09:09

Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody

Published on: September 27, 2024

457
Comparing the Frequency Effect Between the Lexical Decision and Naming Tasks in Chinese
08:08

Comparing the Frequency Effect Between the Lexical Decision and Naming Tasks in Chinese

Published on: April 1, 2016

9.4K

相关实验视频

Last Updated: Jul 4, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

572
Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody
09:09

Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody

Published on: September 27, 2024

457
Comparing the Frequency Effect Between the Lexical Decision and Naming Tasks in Chinese
08:08

Comparing the Frequency Effect Between the Lexical Decision and Naming Tasks in Chinese

Published on: April 1, 2016

9.4K

科学领域:

  • 自然语言处理自然语言处理.
  • 机器学习 机器学习
  • 人工智能的人工智能

背景情况:

  • 跨语言总结 (CLS) 需要理解跨语言的语义和压缩信息.
  • 现有的方法包括使用多语言预训练模型 (mPTM) 的管道 (翻译-然后-总结) 和端到端方法.
  • 由于对资源丰富的数据进行培训,mPTM经常在对低资源语言的语义调整方面扎.

研究的目的:

  • 提出一种新的两阶段微调方法 (TFLCLS),用于低资源的跨语言总结.
  • 增强低资源语言的mPTM的语义对齐和信息压缩能力.
  • 在低资源场景中解决当前CLS方法的局限性.

主要方法:

  • 开发了一种两阶段的微调方法:语义对齐微调,然后是自适应性关节微调.
  • 第1阶段的重点是提高mPTM对低资源语言的理解.
  • 第二阶段增强了CLS任务的语义对齐和信息压缩.

主要成果:

  • 引入了三个新的低资源的CLS数据集:Vi2ZhLow,En2ZhLow和Zh2EnLow.
  • TFLCLS显著超过了最先进的方法,在ROUGE-2中实现了18.88%,12.71%和16.91%的改进.
  • 即使有有限的培训数据 (5,000个样本),也证明了有效的表现.

结论:

  • 拟议的TFLCLS方法有效地解决了低资源跨语言总结的挑战.
  • 两阶段的微调策略增强了关键的语义对齐和信息压缩能力.
  • 在数据稀缺的语言对中,TFLCLS提供了一个有前途的解决方案来提高CLS的性能.