Global speech AI struggles to understand India: Report

Leading systems from OpenAI and Microsoft perform poorly. Sarvam models lead in Indian language recognition. This performance ...
IEEE

Toward Zero-Shot Speech Emotion Recognition Using LLMs in the Absence of Target Data

Abstract: In generalized Speech Emotion Recognition (SER), traditional generalization techniques like transfer learning and domain adaptation rely on access to some amount of unlabeled target domain ...
IEEE

AC-Mix: Self-Supervised Adaptation for Low-Resource Automatic Speech Recognition using Agnostic Contrastive Mixup

Abstract: Self-supervised learning (SSL) leverages large amounts of unlabelled data to learn rich speech representations, fostering improvements in automatic speech recognition (ASR), even when only a ...