During denois- ing, positions whose confidence exceeds 0.9 are committed; if no position exceeds the threshold, the most confident position is committed

We decode candidate lengths within Kpred ±5 , resulting in up to 11 candidates · arXiv 7711.7311

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

representative citing papers

cs.AI · 2026-05-27 · unverdicted · novelty 7.0

DLLM-VSR applies diffusion LLMs to VSR via masked denoising, two-stage training, and length-guided candidate decoding to reach 19.5% WER on LRS3.

Showing 1 of 1 citing paper.

Diffusion Large Language Models for Visual Speech Recognition cs.AI · 2026-05-27 · unverdicted · none · ref 4
DLLM-VSR applies diffusion LLMs to VSR via masked denoising, two-stage training, and length-guided candidate decoding to reach 19.5% WER on LRS3.