Knowledge distillation is an effective way to transfer knowledge from a strong teacher to an efficient student model. Ideally, we expect the better the teacher is, the better the student. However, this expectation does not always come true. It is common that a better teacher model results in a bad student via distillation due to the nonnegligible gap between teacher and student. To bridge the gap, we propose PROD, a PROgressive Distillation method, for dense retrieval. PROD consists of a teacher progressive distillation and a data progressive distillation to gradually improve the student. We conduct extensive experiments on five widely-used benchmarks, MS MARCO Passage, TREC Passage 19, TREC Document 19, MS MARCO Document and Natural Questions, where PROD achieves the state-of-the-art within the distillation methods for dense retrieval. The code and models will be released.
@article{arxiv.2209.13335,
title = {PROD: Progressive Distillation for Dense Retrieval},
author = {Zhenghao Lin and Yeyun Gong and Xiao Liu and Hang Zhang and Chen Lin and Anlei Dong and Jian Jiao and Jingwen Lu and Daxin Jiang and Rangan Majumder and Nan Duan},
journal= {arXiv preprint arXiv:2209.13335},
year = {2023}
}