用重生成概率评估扩散语言模型输出质量，更准更高效
Efficient Self-Evaluation for Diffusion Language Models via Sequence Regeneration
- 通过重生成整个序列的词元概率衡量模型置信度
- 置信度与语义连贯性和答案准确率正相关
- 适合需要自评估和可变长度生成的扩散语言模型应用
扩散大语言模型（dLLMs）因其增强的多样性、可控性和并行性而受到广泛关注。然而，其非序列化、双向掩码生成机制使得质量评估困难，亟需有效的自评估方法。本文提出DiSE，一种简单而高效的dLLMs自评估置信度量化方法。DiSE通过计算在完整上下文条件下重生成整个生成序列中词元的概率来量化置信度，利用词元重生成概率实现更高效可靠的品质评估，支持似然估计与鲁棒不确定性量化。基于DiSE，我们进一步提出一种灵活长度生成框架，根据模型对自身输出的自评估结果自适应控制序列长度。从dLLM泛化性的角度分析并验证了DiSE的可行性，实证表明DiSE与语义连贯性和答案准确性正相关。在似然评估、不确定性量化及灵活长度生成方面的大量实验进一步证实了DiSE的有效性。
原文摘要 · Abstract (English)
Diffusion large language models (dLLMs) have recently attracted significant attention for their ability to enhance diversity, controllability, and parallelism. However, their non-sequential, bidirectionally masked generation makes quality assessment difficult, underscoring the need for effective self-evaluation. In this work, we propose DiSE, a simple yet effective self-evaluation confidence quantification method for dLLMs. DiSE quantifies confidence by computing the probability of regenerating the tokens in the entire generated sequence, given the full context. This method enables more efficient and reliable quality assessment by leveraging token regeneration probabilities, facilitating both likelihood estimation and robust uncertainty quantification. Building upon DiSE, we further introduce a flexible-length generation framework, which adaptively controls the sequence length based on the model's self-assessment of its own output. We analyze and validate the feasibility of DiSE from the perspective of dLLM generalization, and empirically demonstrate that DiSE is positively correlated with both semantic coherence and answer accuracy. Extensive experiments on likelihood evaluation, uncertainty quantification, and flexible-length generation further confirm the effectiveness of the proposed DiSE. Project page: https://zhongzero.github.io/DiSE/
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。