[论文解读] Personalized Dialogue Generation with Diversified Traits

Yinhe Zheng, Guanyi Chen+3 more|arXiv (Cornell University)|Jan 28, 2019

Topic Modeling参考文献 45被引用 89

一句话总结

本文介绍一个具有说话者显式个性特征的大规模数据集，并提出基于特征的 Seq2Seq 模型（PAA 和 PAB）以在多样化特征条件下生成个性化回应。

ABSTRACT

Endowing a dialogue system with particular personality traits is essential to deliver more human-like conversations. However, due to the challenge of embodying personality via language expression and the lack of large-scale persona-labeled dialogue data, this research problem is still far from well-studied. In this paper, we investigate the problem of incorporating explicit personality traits in dialogue generation to deliver personalized dialogues. To this end, firstly, we construct PersonalDialog, a large-scale multi-turn dialogue dataset containing various traits from a large number of speakers. The dataset consists of 20.83M sessions and 56.25M utterances from 8.47M speakers. Each utterance is associated with a speaker who is marked with traits like Age, Gender, Location, Interest Tags, etc. Several anonymization schemes are designed to protect the privacy of each speaker. This large-scale dataset will facilitate not only the study of personalized dialogue generation, but also other researches on sociolinguistics or social science. Secondly, to study how personality traits can be captured and addressed in dialogue generation, we propose persona-aware dialogue generation models within the sequence to sequence learning framework. Explicit personality traits (structured by key-value pairs) are embedded using a trait fusion module. During the decoding process, two techniques, namely persona-aware attention and persona-aware bias, are devised to capture and address trait-related information. Experiments demonstrate that our model is able to address proper traits in different contexts. Case studies also show interesting results for this challenging research problem.

研究动机与目标

• 激发并定义将显式人格特征融入对话生成的任务。
• 提供一个大规模、真实社会对话的数据集，具有多样化特征，适用于可扩展训练。
• 开发融合特征并将其整合到解码过程中的人设感知生成模型。

提出的方法

• 构建 PersonalDialog，一个具有性别、年龄、位置和兴趣标签等特征的大规模中文对话语料库，覆盖 8.47M 说话者和 20.83M 会话。
• 将每个特征编码为嵌入，并与个性特征融合模块合并，形成名为 v_p 的人格表示。
• 实现 v_p 的两种解码集成：(i) Persona-Aware Attention (PAA)，使注意力权重以 v_p 进行条件化；(ii) Persona-Aware Bias (PAB)，通过门控机制在生成分布中加入人格偏置。
• 探索三种特征融合策略：Traits Attention、Traits Average 和 Traits Concatenation。
• 使用一个 Seq2Seq 框架，具有两层 BiGRU 编码器和两层 GRU 解码器，采用 Bahdanau 风格的注意力，基于 v_p 进行条件化。

实验结果

研究问题

• RQ1 显式的人格特征是否能够从大规模社会数据中被有效学习并在生成的对话中表达？
• RQ2 不同的特征融合方法如何影响将人格信息整合到解码器中？
• RQ3 哪种解码策略（PAA 与 PAB）能够最好地利用人格表示来生成与特征一致的回应？
• RQ4 不同的特征融合方案（注意力、平均、连接）在不同情境中对多样化特征表达有何影响？

主要发现

• 该模型能够在不同情境下处理恰当且多样化的特征。
• Persona-Aware Bias (PAB) 通常优于 Persona-Aware Attention (PAA) 在实验中。
• 一个具有真实社会对话和多样化特征的大规模数据集（PersonalDialog）支持个性化对话生成的训练。
• 特征融合使模型能够生成反映显式特征信息的回应，而不要求生成文本包含确切的特征数值。