AI RESEARCH
LESS: Large Language Model Enhanced Semi-Supervised Learning for Speech Foundational Models Using in-the-wild Data
arXiv CS.CL
•
ArXi:2506.04586v3 Announce Type: replace Although state-of-the-art Speech Foundational Models can produce high-quality text pseudo-labels, applying Semi-Supervised Learning (SSL) for in-the-wild real-world data remains challenging due to its richer and complex acoustics compared to curated datasets. To address the challenges, we