Confidence-Aware Self-Distillation for Multimodal Sentiment Analysis with Incomplete Modalities

Yanxi Luo, Shijin Wang, Zhongxing Xu, Yulong Li, Feilong Tang, Jionglong Su · 2025

Read Confidence-Aware Self-Distillation for Multimodal Sentiment Analysis with Incomplete Modalities by Yanxi Luo, Shijin Wang, Zhongxing Xu, Yulong Li, Feilong Tang, Jionglong Su (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-01490.txt.

Description

Luo and five colleagues present a method to judge mood from mixed data when some data are missing in twenty twenty-five. Mood can be read from text, voice, and video at once. In real use, one or more of these streams often drops out. Older fixes rebuild the lost stream or force all streams into one shared form.

The new method scores how sure the model is about each mix of streams. A teacher form of the model then guides a student form to be less unsure. Tests on three standard data sets beat prior methods. The work aims to keep mood tools useful when inputs are patchy.

Catalog record

CREATORYanxi Luo, Shijin Wang, Zhongxing Xu, Yulong Li, Feilong Tang, Jionglong Su
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE30 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-01490

Browse the full library · About the Tower of Babel