A powerful 3B-parameter, LLM-based Reinforcement Learning audio edit model excels at editing emotion, speaking style, and paralinguistics, and features robust zero-shot text-to-speech

TensorFlow implementation of "Multimodal Speech Emotion Recognition using Audio and Text," IEEE SLT-18

Official Repository of Paper: "SynParaSpeech: Automated Synthesis of Paralinguistic Datasets for Speech Generation and Understanding" (ICASSP 2026)

TensorFlow implementation of "Attentive Modality Hopping for Speech Emotion Recognition," ICASSP-20

DiaRemot2-ON: CPU-only audio intelligence pipeline (Faster-Whisper, ONNX, diarization, paralinguistics)

SONATA (SOund and Narrative Advanced Transcription Assistant): An advanced ASR system that captures human expressions including emotive sounds and non-verbal cues.

An ASR model for transcribing laughter and speech-laugh for spontaneous conversational speech

This repo contains the reicpe to assemble a corpus for Foreign Accented English using the crowdsourced corpus Common Voice which contains (optional) accent labels.

The FishBoardMix corpus is designed to explore Speaker-Age estimation technology.

Evaluate language models on syntactic tasks.

Interspeech 2018 Computational Paralinguistics ChallengE (ComParE): Self-Assessed Affect recognition sub-challenge

Machine Learning project in Python (Jupyter Notebook) that detects sarcasm from voice tone, timbre, and intonation, not text. Uses narrowband Mel-spectrograms to capture subtle acoustic and prosodic patterns revealing sarcasm through sound alone.

Experimental framework for systematic analysis of cross-lingual transfer in multilingual speech processing using the Cross-Lingual Transfer Matrix (CLTM).