7 papers
Compress the Cache, Not the Speech Embedding: KV Compression for Efficient Speech LLMs
Ke-Han Lu, Keqi Deng, Ruchao Fan +2
Speech large language models (Speech LLMs) typically encode speech into sequences far longer than text, creating a major efficiency bottleneck during autoregressive decoding. A com…
Rethinking Speech-LLM Integration for ASR: Effective Joint Speech-Text Training by Interleaving
Ruchao Fan, Yiming Wang, Rui Zhao +10
Speech-LLM integration has shown promising results by leveraging extensive textual pretraining, yet its specific benefits for automatic speech recognition (ASR) remain unclear. We…
OpenSTBench: Beyond Semantic Evaluation for Speech Translation
Yanjie An, Yuxiang Zhao, Yichi Zhang +5
Speech translation systems increasingly span speech-to-text translation (S2TT), speech-to-speech translation (S2ST), offline translation, and streaming generation, producing output…
UNIQUE: Universal Top-k Sparse Attention for Training-free Inference and Sparsity-aware Training
Keqi Deng, Shaoshi Ling, Ruchao Fan +1
Long-context inference in large language models (LLMs) is bottlenecked by the linear growth of the self-attention key-value (KV) cache. Top-k sparse attention alleviates this by lo…
Speech LLMs are Contextual Reasoning Transcribers
Keqi Deng, Ruchao Fan, Bo Ren +2
Despite extensions to speech inputs, effectively leveraging the rich knowledge and contextual understanding of large language models (LLMs) in automatic speech recognition (ASR) re…
Multi-head Temporal Latent Attention
Keqi Deng, Philip C. Woodland
While Transformer self-attention offers strong parallelism, the Key-Value (KV) cache grows linearly with sequence length and becomes a bottleneck for inference efficiency. Multi-he…