“别人怎么想？”：基于主观知识的任务导向对话建模
"What do others think?": Task-Oriented Conversational Modeling with Subjective Knowledge
- UNC Chapel Hill（北卡罗来纳大学教堂山分校）
- Amazon, Alexa（亚马逊Alexa）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
针对传统任务导向对话系统难以处理主观请求的问题，提出SK-TOD任务并构建首个含人工标注回复的数据集，揭示多观点聚合等新挑战。
AI中文摘要：
任务导向对话（TOD）系统旨在构建能够帮助用户完成特定目标的对话系统，例如预订酒店或餐厅。传统TOD依赖特定领域的API/DB或外部事实知识来生成回复，无法适应主观的用户请求（例如“WIFI可靠吗？”或“这家餐厅氛围好吗？”）。为解决这一问题，我们提出了一项基于主观知识的TOD（SK-TOD）新任务。我们还提出了首个相应数据集，其中包含寻求主观知识的对话上下文，以及基于主观知识来源人工标注的回复。使用现有TOD方法进行评估时，我们发现该任务带来了新的挑战，例如汇总来自多个知识片段的不同观点。我们希望该任务和数据集能够推动关于TOD及主观内容理解的进一步研究。代码和数据集可在https://github.com/alexa/dstc11-track5获取。
英文摘要：
Task-oriented Dialogue (TOD) Systems aim to build dialogue systems that assist users in accomplishing specific goals, such as booking a hotel or a restaurant. Traditional TODs rely on domain-specific APIs/DBs or external factual knowledge to generate responses, which cannot accommodate subjective user requests (e.g., "Is the WIFI reliable?" or "Does the restaurant have a good atmosphere?"). To address this issue, we propose a novel task of subjective-knowledge-based TOD (SK-TOD). We also propose the first corresponding dataset, which contains subjective knowledge-seeking dialogue contexts and manually annotated responses grounded in subjective knowledge sources. When evaluated with existing TOD approaches, we find that this task poses new challenges such as aggregating diverse opinions from multiple knowledge snippets. We hope this task and dataset can promote further research on TOD and subjective content understanding. The code and the dataset are available at https://github.com/alexa/dstc11-track5.
