DialoGen：面向对话系统的广义长程上下文表示
DialoGen: Generalized Long-Range Context Representation for Dialogue Systems
AI总结：
针对对话中超出最近 k 条话语的长程依赖问题，本文提出 DialoGen 编码器-解码器框架，通过检索最相关历史话语实现紧凑上下文表示，并在 DailyDialog 与 MultiWOZ 上达到接近 SOTA 的效果。
摘要：
长程上下文建模对于对话理解和生成都至关重要。最流行的对话上下文表示方法是按时间顺序拼接最后 k 条话语。然而，对于包含长程依赖关系的对话，即需要查看最后 k 条话语之外的内容才能生成有意义回复时，这种方法可能并不理想。在本文中，我们提出 DialoGen，这是一种新颖的基于编码器-解码器的对话生成框架，具有能够查看最后 k 条话语之外内容的广义上下文表示。我们方法的主要思想是识别并利用最相关的历史话语，而不是最后 k 条，这也使得对话历史能够用更少的 token 进行紧凑表示。我们在对话生成（开放域）和理解（DST）两方面研究了所提方法的有效性。即使采用紧凑的上下文表示，DialoGen 在开放域 DailyDialog 数据集上的表现仍可与最先进模型相媲美。当将所提出的上下文表示应用于现有 DST 模型时，我们在 MultiWOZ 数据集的 DST 任务上观察到类似表现。我们还讨论了 DialoGen 的泛化性和可解释性，并表明先前话语的相关性得分与人类认知高度一致。
英文摘要：
Long-range context modeling is crucial to both dialogue understanding and generation. The most popular method for dialogue context representation is to concatenate the last-$k$ utterances in chronological order. However, this method may not be ideal for conversations containing long-range dependencies, i.e., when there is a need to look beyond last-$k$ utterances to generate a meaningful response. In this work, we propose DialoGen, a novel encoder-decoder based framework for dialogue generation with a generalized context representation that can look beyond the last-$k$ utterances. The main idea of our approach is to identify and utilize the most relevant historical utterances instead of last-$k$, which also enables the compact representation of dialogue history with fewer tokens. We study the effectiveness of our proposed method on both dialogue generation (open-domain) and understanding (DST). Even with a compact context representation, DialoGen performs comparably to the state-of-the-art models on the open-domain DailyDialog dataset. We observe a similar behavior on the DST task of the MultiWOZ dataset when the proposed context representation is applied to existing DST models. We also discuss the generalizability and interpretability of DialoGen and show that the relevance score of previous utterances agrees well with human cognition.
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。
