Back to feed
arXiv cs.CL
arXiv cs.CL
7/9/2026
Audio Sentiment Analysis via Distillation and Cross-Modal Integration of Generated Multilingual Transcripts

Audio Sentiment Analysis via Distillation and Cross-Modal Integration of Generated Multilingual Transcripts

Short summary

A multimodal approach integrates audio features with automatically generated multilingual text transcripts via cascaded cross-modal transformers for speech sentiment analysis. The teacher multimodal model is distilled into a student audio-only model, boosting inference-time performance without computational overhead. Experiments on a large-scale dataset confirm that both automatic transcripts and translations improve sentiment polarity classification, and code is publicly released.

  • •Cross-modal transformers combine audio with ASR-generated multilingual transcripts for sentiment analysis
  • •Knowledge distillation from multimodal teacher to audio-only student improves performance at no inference cost
  • •Code publicly released for reproducibility

Generated with AI, which can make mistakes.

#research-breakthrough#open-source#ai-tools
Read full article at arXiv cs.CL

Is this a good recommendation for you?

Comments

Failed to load comments. Please try again.

Explore more

Continuous Audio Thinking for Large Audio Language Models
arXiv cs.CL
Multilingual Polarization Detection Using Transformer-Based Models with Class Weighting and Threshold Tuning
arXiv cs.CL
Large Language Models as Unified Multimodal Learners for Clinical Prediction
arXiv cs.CL
Deep Temporal Modeling and Ensemble Fusion for Multimodal Emotion Recognition from Physiological Signals
arXiv cs.CL