2024/09/30 by Weiting Tan, Tan, Weiting, Hirofumi Inaguma +7 · 2 citations
Computer Science · #Audio and Speech Processing (eess.AS) #Computation and Language (cs.CL) #FOS: Computer and information sciences #FOS: Electrical engineering #Natural Language Processing Techniques #Sound (cs.SD) #Speech Recognition and Synthesis #Topic Modeling #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2410.00168
openalex publication_date 2024/09/30 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Fusing speech into pre-trained language model (SpeechLM) usually suffers from inefficient encoding of long-form speech and catastrophic forgetting of pre-trained text modality. We propose SSR-Connector (Segmented Speech Representation Connector) for better modality fusion. Leveraging speech-text alignments, our approach segments and compresses speech features to match the granularity of text embeddings. Additionally, we introduce a two-stage training pipeline that includes the distillation and fine-tuning phases to mitigate catastrophic forgetting. SSR-Connector outperforms existing mechanism for speech-text modality fusion, consistently achieving better speech understanding (e.g., +10 accuracy on StoryCloze and +20 on Speech-MMLU) while preserving pre-trained text ability.