4 papers
REVE: Efficient Hallucination Correction for Large Audio-Language Models via Reused Encoder States
Hongjin Song, Jiasheng Kuang, Xinyu Yang +4
Large audio-language models may mention acoustic events that are absent from the input. A separate audio event detector can verify these mentions, but doing so requires a second au…
Rethinking Length-Based Training: Batch Composition and Loss Normalization in Speech Token Language Models
Hongjin Song, Runwu Shi, Weiqiao Shan +4
Short-to-long training is a simple curriculum for speech models, but its gains can be difficult to interpret. In speech token language models, length-based training can change the…
Likelihood-Constrained Acoustic Reranking for Training-Free Hallucination Mitigation in LLM-Based ASR
Jiasheng Kuang, Linru Zheng, Hongjin Song +2
Large language model (LLM)-based automatic speech recognition (ASR) systems achieve strong performance on conventional speech data by leveraging powerful linguistic priors and mult…
Joint Residual Reweighting for Classifier Free Guidance in Flow-Matching Zero-Shot TTS
Runwu Shi, Yujin Wang, Hongjin Song +8
Classifier-free guidance (CFG) is widely used in flow-matching-based zero-shot text-to-speech (TTS), where generation is conditioned on text content and a speech prompt. Standard C…