三思而后行：一种类人两阶段对话智能体用于情感回复生成
Think Twice: A Human-like Two-stage Conversational Agent for Emotional Response Generation
AI总结：
针对情感对话中情感与语义相互制约及标注数据稀缺问题，提出两阶段对话智能体：先由无语料标注的模型生成语义原型回复，再用可控情感精炼器进行修改，在情感生成和语义保持上均优于对比模型。
摘要：
为了实现类人对话系统，当前的情感对话方法使用统一的神经网络联合建模情感与语义。由于情感与语义之间的相互制约，这种策略倾向于生成安全的回复，并且需要稀缺的大规模情感标注对话语料。受人类对话中“三思而后行”行为的启发，我们提出了一种用于情感对话生成的两阶段对话智能体。首先，一个在未使用情感标注对话语料的情况下训练的对话模型生成符合上下文语义的原型回复。其次，第一阶段的原型由一个带有共情假设的可控情感精炼器进行修改。在DailyDialog和EmpatheticDialogues数据集上的实验结果表明，所提出的对话智能体在情感生成方面优于对比模型，并在自动评估和人工评估中保持了语义性能。
英文摘要：
Towards human-like dialogue systems, current emotional dialogue approaches jointly model emotion and semantics with a unified neural network. This strategy tends to generate safe responses due to the mutual restriction between emotion and semantics, and requires rare emotion-annotated large-scale dialogue corpus. Inspired by the "think twice" behavior in human dialogue, we propose a two-stage conversational agent for the generation of emotional dialogue. Firstly, a dialogue model trained without the emotion-annotated dialogue corpus generates a prototype response that meets the contextual semantics. Secondly, the first-stage prototype is modified by a controllable emotion refiner with the empathy hypothesis. Experimental results on the DailyDialog and EmpatheticDialogues datasets demonstrate that the proposed conversational outperforms the comparison models in emotion generation and maintains the semantic performance in automatic and human evaluations.
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。
