Zhen Qin
Research Subjects
Zhen Qin has a total of 21 co-authors. Here are the top co-authors by number of shared publications:
- Xuanhui Wang
Google DeepMind, Mountain View, CA, USA
7 shared publications - Rolf Jagerman
Google Research, Amsterdam, Netherlands
6 shared publications - Michael Bendersky
Google, Mountain View, CA, USA
5 shared publications - Harrie Oosterhuis
University of Amsterdam, Amsterdam, Netherlands
3 shared publications - Honglei Zhuang
Google, Seattle, USA
3 shared publications - Kai Hui
Max Planck Institute for Informatics, Saarbrücken, Germany
2 shared publications
Latest Publications
- Open Access
- 24 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 262-273
Exposure-Based Reinforcement Learning to Rank
Reinforcement learning (RL) methods for learning-to-rank (LTR) can optimize (almost) any ranking goal, e.g., from precision or discounted cumulative gain to fairness-of-exposure or ranking distillation. However, standard RL is ineffective and computationally costly due to the enormous action space in LTR settings. Existing methods reach computational efficiency through custom gradient computation algorithms, but they are very complex to implement and often clash with auto-differentiation. Consequently, existing RL for LTR is not attractive to many practitioners. We reconsider RL for LTR while actively avoiding reliance on custom gradients. Contrary to the existing approaches, we focus on variance reduction and GPU computation. In doing so, we discover that high sample-efficiency can be reached through baseline corrections and partial marginalization. Furthermore, we propose an abstraction that places gradient estimation behind a document-exposure distribution, this enables seamless plug-and-play integration with auto-differentiation. Thereby, one only has to implement a loss as a differentiable function of exposure and RL for LTR can optimize it using auto-differentiation. Our experimental results reveal that our new exposure-based RL for LTR approach converges considerably faster and at significantly higher ranking performance than existing custom gradients, with no additional costs in computation time when using GPUs. In contrast, existing custom gradients result in severe stability issues when converging over many epochs, which never occur for our methods. Thus, we considerably improve RL for LTR methodology by increasing its effectiveness, efficiency, and ease of application.
- 24 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 262-273
Open AccessExposure-Based Reinforcement Learning to Rank
Reinforcement learning (RL) methods for learning-to-rank (LTR) can optimize (almost) any ranking goal, e.g., from precision or discounted cumulative gain to fairness-of-exposure or ranking distillation. However, standard RL is ineffective and computationally costly due to the enormous action space in LTR settings. Existing methods reach computational efficiency through custom gradient computation algorithms, but they are very complex to implement and often clash with auto-differentiation. Consequently, existing RL for LTR is not attractive to many practitioners. We reconsider RL for LTR while actively avoiding reliance on custom gradients. Contrary to the existing approaches, we focus on variance reduction and GPU computation. In doing so, we discover that high sample-efficiency can be reached through baseline corrections and partial marginalization. Furthermore, we propose an abstraction that places gradient estimation behind a document-exposure distribution, this enables seamless plug-and-play integration with auto-differentiation. Thereby, one only has to implement a loss as a differentiable function of exposure and RL for LTR can optimize it using auto-differentiation. Our experimental results reveal that our new exposure-based RL for LTR approach converges considerably faster and at significantly higher ranking performance than existing custom gradients, with no additional costs in computation time when using GPUs. In contrast, existing custom gradients result in severe stability issues when converging over many epochs, which never occur for our methods. Thus, we considerably improve RL for LTR methodology by increasing its effectiveness, efficiency, and ease of application.
- Open Access
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 2361-2371
Optimizing Compound Retrieval Systems
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 2361-2371
Open AccessOptimizing Compound Retrieval Systems
- Open Access
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 3865-3875
Rankers, Judges, and Assistants: Towards Understanding the Interplay of LLMs in Information Retrieval Evaluation
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 3865-3875
Open AccessRankers, Judges, and Assistants: Towards Understanding the Interplay of LLMs in Information Retrieval Evaluation
- Open Access
- 24 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 2307-2317
Reliable Confidence Intervals for Information Retrieval Evaluation Using Generative A.I.
The traditional evaluation of information retrieval (IR) systems is generally very costly as it requires manual relevance annotation from human experts. Recent advancements in generative artificial intelligence -specifically large language models (LLMs)- can generate relevance annotations at an enormous scale with relatively small computational costs. Potentially, this could alleviate the costs traditionally associated with IR evaluation and make it applicable to numerous low-resource applications. However, generated relevance annotations are not immune to (systematic) errors, and as a result, directly using them for evaluation produces unreliable results. In this work, we propose two methods based on prediction-powered inference and conformal risk control that utilize computer-generated relevance annotations to place reliable confidence intervals (CIs) around IR evaluation metrics. Our proposed methods require a small number of reliable annotations from which the methods can statistically analyze the errors in the generated annotations. Using this information, we can place CIs around evaluation metrics with strong theoretical guarantees. Unlike existing approaches, our conformal risk control method is specifically designed for ranking metrics and can vary its CIs per query and document. Our experimental results show that our CIs accurately capture both the variance and bias in evaluation based on LLM annotations, better than the typical empirical bootstrapping estimates. We hope our contributions bring reliable evaluation to the many IR applications where this was traditionally infeasible.
- 24 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 2307-2317
Open AccessReliable Confidence Intervals for Information Retrieval Evaluation Using Generative A.I.
The traditional evaluation of information retrieval (IR) systems is generally very costly as it requires manual relevance annotation from human experts. Recent advancements in generative artificial intelligence -specifically large language models (LLMs)- can generate relevance annotations at an enormous scale with relatively small computational costs. Potentially, this could alleviate the costs traditionally associated with IR evaluation and make it applicable to numerous low-resource applications. However, generated relevance annotations are not immune to (systematic) errors, and as a result, directly using them for evaluation produces unreliable results. In this work, we propose two methods based on prediction-powered inference and conformal risk control that utilize computer-generated relevance annotations to place reliable confidence intervals (CIs) around IR evaluation metrics. Our proposed methods require a small number of reliable annotations from which the methods can statistically analyze the errors in the generated annotations. Using this information, we can place CIs around evaluation metrics with strong theoretical guarantees. Unlike existing approaches, our conformal risk control method is specifically designed for ranking metrics and can vary its CIs per query and document. Our experimental results show that our CIs accurately capture both the variance and bias in evaluation based on LLM annotations, better than the typical empirical bootstrapping estimates. We hope our contributions bring reliable evaluation to the many IR applications where this was traditionally infeasible.
- Open Access
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 2321-2326
Can Query Expansion Improve Generalization of Strong Cross-Encoder Rankers?
Query expansion has been widely used to improve the search results of first-stage retrievers, yet its influence on second-stage, cross-encoder rankers remains under-explored. A recent study shows that current expansion techniques benefit weaker models but harm stronger rankers. In this paper, we re-examine this conclusion and raise the following question: Can query expansion improve generalization of strong cross-encoder rankers? To answer this question, we first apply popular query expansion methods to different cross-encoder rankers and verify the deteriorated zero-shot effectiveness. We identify two vital steps in the experiment: high-quality keyword generation and minimally-disruptive query modification. We show that it is possible to improve the generalization of a strong neural ranker, by generating keywords through a reasoning chain and aggregating the ranking results of each expanded query via self-consistency, reciprocal rank weighting, and fusion. Experiments on BEIR and TREC Deep Learning 2019/2020 show that the nDCG@10 scores of both MonoT5 and RankT5 following these steps are improved, which points out a direction for applying query expansion to strong cross-encoder rankers.
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 2321-2326
Open AccessCan Query Expansion Improve Generalization of Strong Cross-Encoder Rankers?
Query expansion has been widely used to improve the search results of first-stage retrievers, yet its influence on second-stage, cross-encoder rankers remains under-explored. A recent study shows that current expansion techniques benefit weaker models but harm stronger rankers. In this paper, we re-examine this conclusion and raise the following question: Can query expansion improve generalization of strong cross-encoder rankers? To answer this question, we first apply popular query expansion methods to different cross-encoder rankers and verify the deteriorated zero-shot effectiveness. We identify two vital steps in the experiment: high-quality keyword generation and minimally-disruptive query modification. We show that it is possible to improve the generalization of a strong neural ranker, by generating keywords through a reasoning chain and aggregating the ranking results of each expanded query via self-consistency, reciprocal rank weighting, and fusion. Experiments on BEIR and TREC Deep Learning 2019/2020 show that the nDCG@10 scores of both MonoT5 and RankT5 following these steps are improved, which points out a direction for applying query expansion to strong cross-encoder rankers.
- Open Access
- 21 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 4502-4508
Regression Compatible Listwise Objectives for Calibrated Ranking with Binary Relevance
As Learning-to-Rank (LTR) approaches primarily seek to improve ranking quality, their output scores are not scale-calibrated by design. This fundamentally limits LTR usage in score-sensitive applications. Though a simple multi-objective approach that combines a regression and a ranking objective can effectively learn scale-calibrated scores, we argue that the two objectives are not necessarily compatible, which makes the trade-off less ideal for either of them. In this paper, we propose a practical regression compatible ranking (RCR) approach that achieves a better trade-off, where the two ranking and regression components are proved to be mutually aligned. Although the same idea applies to ranking with both binary and graded relevance, we mainly focus on binary labels in this paper. We evaluate the proposed approach on several public LTR benchmarks and show that it consistently achieves either best or competitive result in terms of both regression and ranking metrics, and significantly improves the Pareto frontiers in the context of multi-objective optimization. Furthermore, we evaluated the proposed approach on YouTube Search and found that it not only improved the ranking quality of the production pCTR model, but also brought gains to the click prediction accuracy. The proposed approach has been successfully deployed in the YouTube production system.
- 21 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 4502-4508
Open AccessRegression Compatible Listwise Objectives for Calibrated Ranking with Binary Relevance
As Learning-to-Rank (LTR) approaches primarily seek to improve ranking quality, their output scores are not scale-calibrated by design. This fundamentally limits LTR usage in score-sensitive applications. Though a simple multi-objective approach that combines a regression and a ranking objective can effectively learn scale-calibrated scores, we argue that the two objectives are not necessarily compatible, which makes the trade-off less ideal for either of them. In this paper, we propose a practical regression compatible ranking (RCR) approach that achieves a better trade-off, where the two ranking and regression components are proved to be mutually aligned. Although the same idea applies to ranking with both binary and graded relevance, we mainly focus on binary labels in this paper. We evaluate the proposed approach on several public LTR benchmarks and show that it consistently achieves either best or competitive result in terms of both regression and ranking metrics, and significantly improves the Pareto frontiers in the context of multi-objective optimization. Furthermore, we evaluated the proposed approach on YouTube Search and found that it not only improved the ranking quality of the production pCTR model, but also brought gains to the click prediction accuracy. The proposed approach has been successfully deployed in the YouTube production system.
- Open Access
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 5618-5627
Towards Disentangling Relevance and Bias in Unbiased Learning to Rank
Unbiased learning to rank (ULTR) studies the problem of mitigating various biases from implicit user feedback data such as clicks, and has been receiving considerable attention recently. A popular ULTR approach for real-world applications uses a two-tower architecture, where click modeling is factorized into a relevance tower with regular input features, and a bias tower with bias-relevant inputs such as the position of a document. A successful factorization will allow the relevance tower to be exempt from biases. In this work, we identify a critical issue that existing ULTR methods ignored - the bias tower can be confounded with the relevance tower via the underlying true relevance. In particular, the positions were determined by the logging policy, i.e., the previous production model, which would possess relevance information. We give both theoretical analysis and empirical results to show the negative effects on relevance tower due to such a correlation. We then propose two methods to mitigate the negative confounding effects by better disentangling relevance and bias. Offline empirical results on both controlled public datasets and a large-scale industry dataset show the effectiveness of the proposed approaches. We conduct a live experiment on a popular web store for four weeks, and find a significant improvement in user clicks over the baseline, which ignores the negative confounding effect.
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 5618-5627
Open AccessTowards Disentangling Relevance and Bias in Unbiased Learning to Rank
Unbiased learning to rank (ULTR) studies the problem of mitigating various biases from implicit user feedback data such as clicks, and has been receiving considerable attention recently. A popular ULTR approach for real-world applications uses a two-tower architecture, where click modeling is factorized into a relevance tower with regular input features, and a bias tower with bias-relevant inputs such as the position of a document. A successful factorization will allow the relevance tower to be exempt from biases. In this work, we identify a critical issue that existing ULTR methods ignored - the bias tower can be confounded with the relevance tower via the underlying true relevance. In particular, the positions were determined by the logging policy, i.e., the previous production model, which would possess relevance information. We give both theoretical analysis and empirical results to show the negative effects on relevance tower due to such a correlation. We then propose two methods to mitigate the negative confounding effects by better disentangling relevance and bias. Offline empirical results on both controlled public datasets and a large-scale industry dataset show the effectiveness of the proposed approaches. We conduct a live experiment on a popular web store for four weeks, and find a significant improvement in user clicks over the baseline, which ignores the negative confounding effect.
- Top Cited PapersOpen Access
- 18 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 2308-2313
RankT5: Fine-Tuning T5 for Text Ranking with Ranking Losses
Pretrained language models such as BERT have been shown to be exceptionally effective for text ranking. However, there are limited studies on how to leverage more powerful sequence-to-sequence models such as T5. Existing attempts usually formulate text ranking as a classification problem and rely on postprocessing to obtain a ranked list. In this paper, we propose RankT5 and study two T5-based ranking model structures, an encoder-decoder and an encoder-only one, so that they not only can directly output ranking scores for each query-document pair, but also can be fine-tuned with pairwise or listwise ranking losses to optimize ranking performance. Our experiments show that the proposed models with ranking losses can achieve substantial ranking performance gains on different public text ranking data sets. Moreover, ranking models fine-tuned with listwise ranking losses have better zero-shot ranking performance on out-of-domain data than models fine-tuned with classification losses.
- 18 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 2308-2313
Top Cited PapersOpen AccessRankT5: Fine-Tuning T5 for Text Ranking with Ranking Losses
Pretrained language models such as BERT have been shown to be exceptionally effective for text ranking. However, there are limited studies on how to leverage more powerful sequence-to-sequence models such as T5. Existing attempts usually formulate text ranking as a classification problem and rely on postprocessing to obtain a ranked list. In this paper, we propose RankT5 and study two T5-based ranking model structures, an encoder-decoder and an encoder-only one, so that they not only can directly output ranking scores for each query-document pair, but also can be fine-tuned with pairwise or listwise ranking losses to optimize ranking performance. Our experiments show that the proposed models with ranking losses can achieve substantial ranking performance gains on different public text ranking data sets. Moreover, ranking models fine-tuned with listwise ranking losses have better zero-shot ranking performance on out-of-domain data than models fine-tuned with classification losses.
- Open Access
- 25 August 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 26 (12) , 5680-5691
Event Detection in Continuous Video: An Inference in Point Process Approach
We propose a novel approach toward event detection in real-world continuous video sequences. The method: 1) is able to model arbitrary-order non-Markovian dependences in videos to mitigate local visual ambiguities; 2) conducts simultaneous event segmentation and labeling; and 3) is time-window free. The idea is to represent a video as an event stream of both high-level semantic events and low-level video observations. In training, we learn a point process model called a piecewise-constant conditional intensity model (PCIM) that is able to capture complex non-Markovian dependences in the event streams. In testing, event detection can be modeled as the inference of high-level semantic events, given low-level image observations. We develop the first inference algorithm for PCIM and show it samples exactly from the posterior distribution. We then evaluate the video event detection task on real-world video sequences. Our model not only provides competitive results on the video event segmentation and labeling task, but also provides benefits, including being interpretable and efficient.
- 25 August 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 26 (12) , 5680-5691
Open AccessEvent Detection in Continuous Video: An Inference in Point Process Approach
We propose a novel approach toward event detection in real-world continuous video sequences. The method: 1) is able to model arbitrary-order non-Markovian dependences in videos to mitigate local visual ambiguities; 2) conducts simultaneous event segmentation and labeling; and 3) is time-window free. The idea is to represent a video as an event stream of both high-level semantic events and low-level video observations. In training, we learn a point process model called a piecewise-constant conditional intensity model (PCIM) that is able to capture complex non-Markovian dependences in the event streams. In testing, event detection can be modeled as the inference of high-level semantic events, given low-level image observations. We develop the first inference algorithm for PCIM and show it samples exactly from the posterior distribution. We then evaluate the video event detection task on real-world video sequences. Our model not only provides competitive results on the video event segmentation and labeling task, but also provides benefits, including being interpretable and efficient.