VCRS benchmark datasets

VCRS基准数据集是由新加坡科技研究局高性能计算研究所的研究团队创建，专注于电子商务和电影领域的语音对话推荐系统。数据集通过ChatGPT生成的对话模板和神经语音合成技术，将用户-物品交互转换为自然的语音对话。创建过程包括数据选择、文本对话生成、语音对话生成和质量评估。该数据集旨在解决传统文本对话推荐系统在用户体验和可访问性方面的问题，特别适用于视觉障碍或阅读写作能力有限的用户。

The VCRS benchmark dataset was developed by a research team from the Institute of High Performance Computing, Agency for Science, Technology and Research (A*STAR) of Singapore, focusing on spoken dialogue recommendation systems in the e-commerce and film domains. This dataset converts user-item interactions into natural spoken dialogues using dialogue templates generated by ChatGPT and neural speech synthesis techniques. Its construction process includes data selection, textual dialogue generation, spoken dialogue generation, and quality assessment. This dataset aims to address the issues of traditional textual dialogue recommendation systems in terms of user experience and accessibility, and is particularly suitable for users with visual impairments or limited reading and writing abilities.