Length-Adaptive Decoding for Masked Diffusion Machine Translation

• 类型：arxiv
• 标识：2608.22274
• 链接：https://arxiv.org/abs/2608.22274
• 主分类：multimodal
• 形态：method
• TLDR：Machine translation tests masked diffusion language models (dLLMs) because every source token must be rendered faithfully, while fixed canvas decoding must choose target length before denoising. Existing masked diffusion decoding work mainly studies token unmasking order, leaving this length decision under-explored despite its direct effect on coverage and redundancy. We introduce Entropy-Valley (EV), a training-free length selector that scores candidate target canvases by mean predictive entropy from all-mask forward passes and selects the canvas the backbone is most prepared to fill. Relativ
• 待LLM分类：否
• 来源文件：
• /inbox/tom/_candidates/2026-08-26-agent-rag-longcontext-candidates.json