Sheng Li

Cached LLM Probability Retrieval for Speech Recognition

Large language models (LLMs) enhance automatic speech recognition (ASR) by providing linguistic priors; however, their direct rescoring is costly because it requires evaluating every N-best hypothesis. This paper introduces"cached LLM probability retrieval,"which involves querying a local teacher LLM offline to obtain...

Stacking Complementary CLAP Embeddings for Improving Text-Audio Alignment Correspondence Scoring

Agentic-DuplexGen: Decoupling Content, Timing, and Acoustics for Synthetic Dialogue Speech

Synthetic conversational speech has become an important resource for developing and evaluating conversational speech systems. However, existing dialogue synthesis pipelines typically generate dialogue content first and then insert interruptions, overlap, and backchannels using handcrafted markers or timing rules, makin...

DuplexGen: Decoupling Content, Timing, and Acoustics for Synthetic Dialogue Speech