paper-with-me

Papers

Whisfusion: Parallel ASR Decoding with Masked Diffusion

2025-08-09 · Taeyoun Kwon, Junhyuk Ahn, Taegeun Yun, Heeju Jwa, Yoonchae Choi, Siwon Park, Jongchan Kim, Hyungon Ryu, Hyuk-Jae Lee, Nam-Joon Kim arxiv

Autoregressive (AR) encoder-decoder models dominate high-quality multilingual ASR, but their left-to-right decoders make inference latency scale with transcript length. A natural alternative, CTC-style non-autoregressive (NAR) systems avoid this bottleneck but their conditional independence assumption sacrifices transcript-level generative modeling. Masked diffusion language models (e.g., LLaDA, MDLM) offer a competitive NAR text-generation approach. We ask whether such models can bring NAR ASR into the accuracy regime of strong AR ASR systems while removing the left-to-right bottleneck. We propose Whisfusion, which trains a dedicated masked diffusion decoder from scratch on top of frozen Whisper-large-v3 audio embeddings, denoising masked transcripts in just a few steps. We train on ~68k hours of 11-language speech with high-mask specialization to align training with the fully masked starting point of inference, and decode via Parallel Diffusion Decoding. Whisfusion surpasses Whisper-large-v3 on group-average accuracy across English, European, and CJK benchmarks, while running 4-5x faster, additionally surpassing Whisper-turbo in both accuracy and throughput. It reaches accuracy competitive with Canary and Qwen3-ASR while running 3-7x faster. These results establish masked diffusion as a Pareto-competitive non-autoregressive paradigm for high-throughput multilingual transcription. Code and model weights are available at https://github.com/taeyoun811/Whisfusion.

📄 PDF Abstract BibTeX arXiv:2508.07048

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Supportive Token Revealing for Fast Diffusion Language Model Decoding

2026-06-02 · Giries Abu Ayoub, Mario Barbara, Lluís Pastor-Pérez, Tanja Bien 외 arxiv

Discrete diffusion language models can generate text efficiently by updating multiple masked positions in parallel, but this parallelism introduces a quality-latency trade-off. Aggressive decoding may commit mutually dep…

Cluster-Level Attention-Guided Parallel Decoding for Masked Diffusion Language Models

2026-05-28 · Heqiang Qi, Wei Huang, Mingyuan Bai, Xiangming Meng arxiv

Masked diffusion language models (MDLMs) enable parallel decoding by predicting all masked positions at each denoising step, yet existing training-free samplers usually decide which positions to commit at token-level gra…

Divide and Conquer: Accelerating Diffusion-Based Large Language Models via Adaptive Parallel Decoding

2026-02-27 · Xiangzhong Luo, Yilin An, Zhicheng Yu, Weichen Liu 외 arxiv

Diffusion-based large language models (dLLMs) have shown promising performance across various reasoning tasks, establishing themselves as an alternative to autoregressive large language models (LLMs). Unlike autoregressi…

Auto-Regressive Masked Diffusion Models

2026-01-23 · Mahdi Karami, Ali Ghodsi arxiv

Masked diffusion models (MDMs) have emerged as a promising approach for language modeling, yet they face a performance gap compared to autoregressive models (ARMs) and require more training iterations. In this work, we p…

Text Generation

DOS: Dependency-Oriented Sampler for Masked Diffusion Language Models

2026-03-16 · Xueyu Zhou, Yangrong Hu, Jian Huang arxiv

Masked diffusion language models (MDLMs) have recently emerged as a new paradigm in language modeling, offering flexible generation dynamics and enabling efficient parallel decoding. However, existing decoding strategies…

Mathematical ReasoningCode Generation