logo
CanyamAI summaries for academic research
language
MobileMobile
  • Phone FeatureWeChat Mini Program
EnglishEnglish
noticeNotice
BlogBlog
delete
arrow
Return

Entropy-Guided Meta-Initialization regularization for few-shot text classification

delete2025-07-02
delete0
PRE
AI
J
J. Shin
J
Jinwoo Kim
J
Jangho Kim
DOI：10.1016/j.knosys.2025.113961delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
• We define blind class over-confidence as some classes’ overconfidence before adaptation. • We propose a gradient-based method that alleviates blind over-confidence via entropy. • Our method improves Clinc150 1/5-shot by 5.86%/7.23% with better meta-initialization.

Journal

K
Knowledge-Based Systems
IF:
7.6
Papers:
1.2W
Citations:
4.5W

Organization

No organization information available