Abstract | ||
---|---|---|
Two-tower architecture is commonly used in real-world systems for Unbiased Learning to Rank (ULTR), where a Deep Neural Network (DNN) tower models unbiased relevance predictions, while another tower models observation biases inherent in the training data like user clicks. This two-tower architecture introduces inductive biases to allow more efficient use of limited observational logs and better generalization during deployment than single-tower architecture that may learn spurious correlations between relevance predictions and biases. However, despite their popularity, it is largely neglected in the literature that existing two-tower models assume that the joint distribution of relevance prediction and observation probabilities are completely factorizable. In this work, we revisit two-tower models for ULTR. We rigorously show that the factorization assumption can be too strong for real-world user behaviors, and existing methods may easily fail under slightly milder assumptions. We then propose several novel ideas that consider a wider spectrum of user behaviors while still under the two-tower framework to maintain simplicity and generalizability. Our concerns of existing two-tower models and the effectiveness of our proposed methods are validated on both controlled synthetic and large-scale real-world datasets. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1145/3477495.3531837 | SIGIR '22: Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval |
Keywords | DocType | Citations |
Unbiased Learning to Rank, Expectation Maximization, Bias Factorization | Conference | 1 |
PageRank | References | Authors |
0.35 | 19 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Le Yan | 1 | 1 | 0.35 |
Zhen Qin | 2 | 138 | 16.93 |
Honglei Zhuang | 3 | 193 | 16.37 |
Xuanhui Wang | 4 | 1394 | 68.85 |
Michael Bendersky | 5 | 986 | 48.69 |
Marc A. Najork | 6 | 2538 | 278.16 |