arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.03769cs.LG
收藏

在扩散语言模型中使用潜在令牌进行推理

Reasoning with Latent Tokens in Diffusion Language Models

Carnegie Mellon University(卡内基梅隆大学)

Andre He, Sean Welleck, Daniel Fried

AI总结:

本文提出通过引入潜在令牌提升扩散语言模型在需要全局一致性和前瞻性的推理任务中的性能。

摘要:

离散扩散模型最近在语言建模中与自回归模型竞争,甚至在需要规划和全局一致性推理任务上超越了自回归模型,但它们在推理时需要更多的计算。我们追溯这种权衡到一个关键机制:扩散模型被训练以联合预测所有未知令牌的分布,包括那些在当前步骤不会被解码的令牌。去除这种联合预测会加快推理但会降低性能,这表明在解码位置的准确预测依赖于对未解码令牌分布的联合推理。我们将这些视为潜在令牌,并引入一种方法来调节其数量,实验证明这能够实现推理速度和样本质量之间的平滑权衡。此外,我们证明可以通过辅助的多令牌预测目标将潜在令牌引入自回归模型,在传统上挣扎的相同推理任务上取得了显著改进。我们的结果表明,尽管潜在令牌在扩散模型中自然出现,但它们代表了一种通用机制,可用于提高需要全局一致性和前瞻性的任务性能。

英文摘要:

Discrete diffusion models have recently become competitive with autoregressive models for language modeling, even outperforming them on reasoning tasks requiring planning and global coherence, but they require more computation at inference time. We trace this trade-off to a key mechanism: diffusion models are trained to jointly predict a distribution over all unknown tokens, including those that will not actually be decoded in the current step. Ablating this joint prediction yields faster inference but degrades performance, revealing that accurate prediction at the decoded position relies on joint reasoning about the distribution of undecoded tokens. We interpret these as latent tokens and introduce a method for modulating their number, demonstrating empirically that this enables a smooth tradeoff between inference speed and sample quality. Furthermore, we demonstrate that latent tokens can be introduced into autoregressive models through an auxiliary multi-token prediction objective, yielding substantial improvements on the same reasoning tasks where they have traditionally struggled. Our results suggest that latent tokens, while arising naturally in diffusion, represent a general mechanism for improving performance on tasks requiring global coherence or lookahead.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑