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.
@article{arxiv.2410.00168,
title = {SSR: Alignment-Aware Modality Connector for Speech Language Models},
author = {Weiting Tan and Hirofumi Inaguma and Ning Dong and Paden Tomasello and Xutai Ma},
journal= {arXiv preprint arXiv:2410.00168},
year = {2025}
}