| title | An Empirical Study on Zero-Data Bootstrapping for Conversational Recommender Systems |
|---|---|
| authors | Rohan Surana, Junda Wu, Zhouhang Xie, Yu Xia, Nathan Kallus, Julian McAuley |
| date | 2026-08-28 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2504.15476 |
| affiliation | University of California at San Diego |
| upvotes | 1 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | Conversational Recommender Systems (CRS) typically require domain-specific dialogue data, which is costly, scarce, and often unavailable in new domains. We conduct a systematic empirical study of zero-data CRS bootstrapping: generating synthetic conversational supervision from non-conversational signals---item reviews, metadata, and user-item interactions---without any in-domain dialogue corpus. We compare two information-theoretic selection strategies, Jensen-Shannon diversity and Fisher information, across domain signals, model architectures, datasets, and fine-tuning paradigms. Our results show that domain-grounded synthetic data consistently outperforms zero-shot prompting and naive synthetic baselines; active selection improves data efficiency over random sampling; metadata and collaborative filtering signals each improve selection quality; and, in low-resource settings, synthetic data can outperform scarce real dialogues while further complementing them. These findings establish non-conversational domain signals as a viable path toward building CRS without conversational training data. The code is available at https://anonymous.4open.science/r/zero_data_crs/ . |
cite
@misc{surana2026empirical,
title={An Empirical Study on Zero-Data Bootstrapping for Conversational Recommender Systems},
author={Rohan Surana and Junda Wu and Zhouhang Xie and Yu Xia and Nathan Kallus and Julian McAuley},
year={2026},
eprint={2504.15476},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2504.15476}
}