Controlling Dialogue Generation with Semantic Exemplars
Controlling Dialogue Generation with Semantic Exemplars
复制标题
DOI：
10.18653/v1/2021.naacl-main.240
复制
发表时间：
2020-08
期刊：
影响因子：
--
通讯作者：
Prakhar Gupta;Jeffrey P. Bigham;Yulia Tsvetkov;Amy Pavel
中科院分区：
文献类型：
--
作者：
Prakhar Gupta;Jeffrey P. Bigham;Yulia Tsvetkov;Amy Pavel
用大型语言模型预先训练的对话系统会生成本地连贯的响应，但缺乏对实现特定目标所需的响应的细粒度控制。一种有前途的方法来控制响应生成是基于范例的生成，其中模型编辑从训练数据中检索的范例响应，或手写以战略性地解决话语级目标，以适应新的对话上下文。我们提出了一个基于范例的对话生成模型，EDGE，它使用的语义框架中存在的范例响应，以指导响应生成。我们表明，控制对话生成的基础上的语义框架的范例，提高了生成的响应的一致性，同时保持语义意义和会话目标的范例响应。
Dialogue systems pretrained with large language models generate locally coherent responses, but lack fine-grained control over responses necessary to achieve specific goals. A promising method to control response generation is exemplar-based generation, in which models edit exemplar responses that are retrieved from training data, or hand-written to strategically address discourse-level goals, to fit new dialogue contexts. We present an Exemplar-based Dialogue Generation model, EDGE, that uses the semantic frames present in exemplar responses to guide response generation. We show that controlling dialogue generation based on the semantic frames of exemplars improves the coherence of generated responses, while preserving semantic meaning and conversation goals present in exemplar responses.