🧪 Test?View on arXiv
Incongruence-Aware Evaluation of Multimodal Speech Emotion Recognition in Audio Large Language Models
Author1, Author2, Author3, Author4, Author5
multimodalspeech emotion recognitionlarge language models
2610.40883
Builder Relevance
1d ago70%
Abstract
This paper investigates the impact of incongruence between spoken content and vocal delivery on speech emotion recognition in large language models.
Reality Card
Core Claim
Incongruence between audio and text channels significantly increases error rates in speech emotion recognition across multiple datasets.
Method / Result
The incongruence score predicts error in all three evaluated corpora, outperforming six simpler alternatives.
Limitations
The results may vary significantly based on the choice of model for text processing, indicating potential instability in findings.
Paper to code
Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.
No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.