Name
Title | ||
|---|---|---|
P3 Ranker: Mitigating the Gaps between Pre-training and Ranking Fine-tuning with Prompt-based Learning and Pre-finetuning |
Abstract | ||
|---|---|---|
Compared to other language tasks, applying pre-trained language models (PLMs) for search ranking often requires more nuances and training signals. In this paper, we identify and study the two mismatches between pre-training and ranking fine-tuning: the training schema gap regarding the differences in training objectives and model architectures, and the task knowledge gap considering the discrepancy between the knowledge needed in ranking and that learned during pre-training. To mitigate these gaps, we propose Pre-trained, Prompt-learned and Pre-finetuned Neural Ranker (P3 Ranker). P3 Ranker leverages prompt-based learning to convert the ranking task into a pre-training like schema and uses pre-finetuning to initialize the model on intermediate supervised tasks. Experiments on MS MARCO and Robust04 show the superior performances of P3 Ranker in few-shot ranking. Analyses reveal that P3 Ranker is able to better accustom to the ranking task through prompt-based learning and retrieve necessary ranking-oriented knowledge gleaned in pre-finetuning, resulting in data-efficient PLM adaptation. Our code is available at https://github.com/NEUIR/P3Ranker. |
Year | DOI | Venue |
|---|---|---|
2022 | 10.1145/3477495.3531786 | SIGIR '22: Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval |
Keywords | DocType | Citations |
Prompt-based Learning, Pre-finetuning, Few-shot Ranking | Conference | 0 |
PageRank | References | Authors |
0.34 | 8 | 6 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Xiaomeng Hu | 1 | 0 | 0.34 |
| Shi Yu | 2 | 8 | 1.21 |
| Chen-Yan Xiong | 3 | 405 | 30.82 |
| Zheng-Hao Liu | 4 | 9 | 3.20 |
| Zhiyuan Liu | 5 | 2037 | 123.68 |
| Ge YU | 6 | 1313 | 175.88 |