paper-with-me

Papers

PART: Progressive Alignment Representation Training for Multilingual Speech-To-Text with LLMs

2025-09-24 · Pei Zhang, Andong Chen, Xi Chen, Baosong Yang, Derek F. Wong, Fei Huang arxiv

Large language models (LLMs) have expanded from text to speech, giving rise to Speech Large Models (SLMs) that support recognition, translation, and synthesis. A key challenge is aligning speech and text representations, which becomes harder in multilingual settings. Existing methods often freeze LLM parameters and train encoders on multilingual data, but this forces cross-language convergence and limits performance. We introduce Progressive Alignment Representation Training (PART), a multi-stage and multi-task framework that separates within-language from cross-language alignment. During cross-language training, LLM parameters are dynamically activated, and text-based tasks are later introduced to enhance multilingual understanding. Experiments on CommonVoice 15, Fleurs, Wenetspeech, and CoVoST2 show that PART surpasses conventional approaches, with analysis confirming its ability to balance language-specific distinctions and cross-language generalization. These results demonstrate PART's effectiveness and generality for multilingual speech modality alignment.

📄 PDF Abstract BibTeX arXiv:2509.19745

Code (0)

등록된 구현이 없습니다.

Tasks

Text to Speech

Similar Papers 제목 키워드 기반

A Progressive Framework of Vision-language Knowledge Distillation and Alignment for Multilingual Scene

2024-04-17 · Wenbo Zhang, Yifan Zhang, Jianfeng Lin, Binqiang Huang 외

Pre-trained vision-language (V-L) models such as CLIP have shown excellent performance in many downstream cross-modal tasks. However, most of them are only applicable to the English context. Subsequent research has focus…

image-classificationImage ClassificationKnowledge DistillationLanguage Modelling+1

ProCLIP: Progressive Vision-Language Alignment via LLM-based Embedder

2025-10-21 · Xiaoxing Hu, Kaicheng Yang, Ziyang Gong, Qi Ming 외 arxiv

The original CLIP text encoder is limited by a maximum input length of 77 tokens, which hampers its ability to effectively process long texts and perform fine-grained semantic understanding. In addition, the CLIP text en…

Contrastive Learning

High-Dimensional Interlingual Representations of Large Language Models

2025-03-14 · Bryan Wilie, Samuel Cahyawijaya, Junxian He, Pascale Fung

Large language models (LLMs) trained on massive multilingual datasets hint at the formation of interlingual constructs--a shared subspace in the representation space. However, evidence regarding this phenomenon is mixed,…

TokAlign++: Advancing Vocabulary Adaptation via Better Token Alignment

2026-05-13 · Chong Li, Yingzhuo Deng, Wen Yang, Jiajun Zhang 외 arxiv

Tokenization is a foundational step in the text process of Large Language Models (LLMs). Texts must be first tokenized into token IDs, which are then input to LLMs. Inefficient tokenization results in long token-ID seque…

PreAlign: Boosting Cross-Lingual Transfer by Early Establishment of Multilingual Alignment

2024-07-23 · Jiahuan Li, ShuJian Huang, Aarron Ching, Xinyu Dai 외

Large language models demonstrate reasonable multilingual abilities, despite predominantly English-centric pretraining. However, the spontaneous multilingual alignment in these models is shown to be weak, leading to unsa…

Cross-Lingual TransferLanguage ModelingLanguage ModellingZero-Shot Cross-Lingual Transfer