A local pipeline makes speech-emotion evidence inspectable and answerable by a small LLM panel while staying on-device, but zero-shot emotion accuracy is poor and CPU runtime exceeds real time.
arXiv preprint arXiv:2508.14130 (2025)
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
representative citing papers
Lasso-selected speech tokens enhance text LLMs for multimodal classification by reducing long audio sequences to task-relevant features via self-supervised adaptation.
citing papers explorer
-
EmotionAI: A Privacy-Preserving Computational Intelligence Pipeline for Speech-Emotion-Grounded Conversational Analysis
A local pipeline makes speech-emotion evidence inspectable and answerable by a small LLM panel while staying on-device, but zero-shot emotion accuracy is poor and CPU runtime exceeds real time.
-
A Simple Method to Enhance Pre-trained Language Models with Speech Tokens for Classification
Lasso-selected speech tokens enhance text LLMs for multimodal classification by reducing long audio sequences to task-relevant features via self-supervised adaptation.