Vogue
收藏资源简介:
Vogue是一个为时尚购物场景下的对话式推荐而设计的新型多模态数据集,包含60个真实的人与人之间的对话,每个对话都配有一个共享的视觉目录、商品元数据、用户时尚配置文件/历史以及来自寻求者和助手双方的对话后评分。该数据集旨在解决现有对话式推荐数据集在模拟对话、缺少用户历史和详细反馈等方面的局限性,为研究多模态对话系统提供了独特的资源。
Vogue is a novel multimodal dataset designed for conversational recommendation in fashion shopping scenarios. It contains 60 real human-to-human conversations, each paired with a shared visual catalog, item metadata, user fashion profiles/histories, and post-conversation ratings from both seekers and assistants. This dataset aims to address the limitations of existing conversational recommendation datasets, including simulated dialogues, insufficient user historical data and lack of detailed feedback, providing a unique resource for research on multimodal conversational systems.

- 1VOGUE: A Multimodal Dataset for Conversational Recommendation in FashionUniversity of Toronto, University of Waterloo · 2025年



