Xiang Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Xiang Li has a total of 57 co-authors. Here are the top co-authors by number of shared publications:
- Wei Lin
Meituan, Beijing, China
14 shared publications - Ruidong Han
Meituan, Beijing, China
9 shared publications - Fei Jiang
Meituan, Beijing, China
7 shared publications - Bin Yin
Meituan, Beijing, China
5 shared publications - Fuzhen Zhuang
School of Artificial Intelligence, Beihang University, Beijing, China
3 shared publications - Junjie Xie
Waimai, Meituan, China
3 shared publications
Latest Publications
- Open Access
- 19 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
An Item-side Competitive Fairness for Recommender Systems
The ranking algorithm of the recommender system would influence the utilities of both the users and the item producers. An algorithm that only optimizes users’ utility may lead to an unfair distribution of exposure over the items and further hurt the utility of the corresponding item producers. To address this issue, researchers and practitioners have recently proposed ensuring a fair exposure distribution among all the items. However, we argue that the ultimate goal of item-side fairness is not to ensure any items would receive exposure equally or proportional to their merit but to foster fair competition among the producers on the platform. Therefore, this paper introduces a novel item-side individual fairness named competitive fairness based on the competitive relationship between item producers. Competitive fairness argues that the key to fostering fair market competition resides in ensuring exposure equity among items with competitive relationships rather than items without competitive relationships.We further define a competitive fairness metric based on these assumptions and propose a G radient-based R e-ranking algorithm for C ompetitive F airness (GRCF) to optimize the proposed competitive fairness metric while maintaining good ranking quality. We conducted a series of simulation experiments to investigate how various competitive relationships impact our algorithm's performance. Experiments on several real-world datasets with genuine competitive relationships verify that our approach can achieve superior performance compared to the baselines and that the effectiveness of our algorithm is independent of the specific recommendation algorithm used.
- 19 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
Open AccessAn Item-side Competitive Fairness for Recommender Systems
The ranking algorithm of the recommender system would influence the utilities of both the users and the item producers. An algorithm that only optimizes users’ utility may lead to an unfair distribution of exposure over the items and further hurt the utility of the corresponding item producers. To address this issue, researchers and practitioners have recently proposed ensuring a fair exposure distribution among all the items. However, we argue that the ultimate goal of item-side fairness is not to ensure any items would receive exposure equally or proportional to their merit but to foster fair competition among the producers on the platform. Therefore, this paper introduces a novel item-side individual fairness named competitive fairness based on the competitive relationship between item producers. Competitive fairness argues that the key to fostering fair market competition resides in ensuring exposure equity among items with competitive relationships rather than items without competitive relationships.We further define a competitive fairness metric based on these assumptions and propose a G radient-based R e-ranking algorithm for C ompetitive F airness (GRCF) to optimize the proposed competitive fairness metric while maintaining good ranking quality. We conducted a series of simulation experiments to investigate how various competitive relationships impact our algorithm's performance. Experiments on several real-world datasets with genuine competitive relationships verify that our approach can achieve superior performance compared to the baselines and that the effectiveness of our algorithm is independent of the specific recommendation algorithm used.
- Open Access
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2482-2493
MTGenRec: An Efficient Distributed Training System for Generative Recommendation Models in Meituan
Recommendation is crucial for both user experience and company revenue in Meituan as a leading lifestyle company, and generative recommendation models (GRMs) are shown to produce quality recommendations recently. However, existing systems are limited by insufficient functionality support and inefficient implementations for training GRMs in industrial scenarios. As such, we introduce MTGenRec as an efficient and scalable system for GRM training. Specifically, to handle real-time insertions/deletions of sparse embeddings, MTGenRec employs dynamic hash tables to replace static ones. To improve training efficiency, MTGenRec conducts dynamic sequence balancing to address the computation load imbalances among GPUs and adopts feature ID deduplication alongside automatic table merging to accelerate embedding lookup. Extensive experiments show that MTGenRec improves training throughput by 1.6x - 2.4x while achieving good scalability when running over 100 GPUs. MTGenRec has been deployed for many applications in Meituan and is now handling hundreds of millions of requests on a daily basis. On the delivery platform, we observe a 1.22% growth in user order volume and a 1.31% enhancement in online PV_CTR.
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2482-2493
Open AccessMTGenRec: An Efficient Distributed Training System for Generative Recommendation Models in Meituan
Recommendation is crucial for both user experience and company revenue in Meituan as a leading lifestyle company, and generative recommendation models (GRMs) are shown to produce quality recommendations recently. However, existing systems are limited by insufficient functionality support and inefficient implementations for training GRMs in industrial scenarios. As such, we introduce MTGenRec as an efficient and scalable system for GRM training. Specifically, to handle real-time insertions/deletions of sparse embeddings, MTGenRec employs dynamic hash tables to replace static ones. To improve training efficiency, MTGenRec conducts dynamic sequence balancing to address the computation load imbalances among GPUs and adopts feature ID deduplication alongside automatic table merging to accelerate embedding lookup. Extensive experiments show that MTGenRec improves training throughput by 1.6x - 2.4x while achieving good scalability when running over 100 GPUs. MTGenRec has been deployed for many applications in Meituan and is now handling hundreds of millions of requests on a daily basis. On the delivery platform, we observe a 1.22% growth in user order volume and a 1.31% enhancement in online PV_CTR.
- Open Access
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 5366-5370
A Soft-partitioned Semi-supervised Collaborative Transfer Learning Approach for Multi-Domain Recommendation
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 5366-5370
Open AccessA Soft-partitioned Semi-supervised Collaborative Transfer Learning Approach for Multi-Domain Recommendation
- Top Cited PapersOpen Access
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 5731-5738
MTGR: Industrial-Scale Generative Recommendation Framework in Meituan
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 5731-5738
Top Cited PapersOpen AccessMTGR: Industrial-Scale Generative Recommendation Framework in Meituan
- Open Access
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 3777-3781
Enhancing CTR Prediction through Sequential Recommendation Pre-training: Introducing the SRP4CTR framework
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 3777-3781
Open AccessEnhancing CTR Prediction through Sequential Recommendation Pre-training: Introducing the SRP4CTR framework
- Open Access
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 3374-3383
Aligning Explanations for Recommendation with Rating and Feature via Maximizing Mutual Information
Providing natural language-based explanations to justify recommendations helps to improve users' satisfaction and gain users' trust. However, as current explanation generation methods are commonly trained with an objective to mimic existing user reviews, the generated explanations are often not aligned with the predicted ratings or some important features of the recommended items, and thus, are suboptimal in helping users make informed decision on the recommendation platform. To tackle this problem, we propose a flexible model-agnostic method named MMI (Maximizing Mutual Information) framework to enhance the alignment between the generated natural language explanations and the predicted rating/important item features. Specifically, we propose to use mutual information (MI) as a measure for the alignment and train a neural MI estimator. Then, we treat a well-trained explanation generation model as the backbone model and further fine-tune it through reinforcement learning with guidance from the MI estimator, which rewards a generated explanation that is more aligned with the predicted rating or a pre-defined feature of the recommended item. Experiments on three datasets demonstrate that our MMI framework can boost different backbone models, enabling them to outperform existing baselines in terms of alignment with predicted ratings and item features. Additionally, user studies verify that MI-enhanced explanations indeed facilitate users' decisions and are favorable compared with other baselines due to their better alignment properties.
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 3374-3383
Open AccessAligning Explanations for Recommendation with Rating and Feature via Maximizing Mutual Information
Providing natural language-based explanations to justify recommendations helps to improve users' satisfaction and gain users' trust. However, as current explanation generation methods are commonly trained with an objective to mimic existing user reviews, the generated explanations are often not aligned with the predicted ratings or some important features of the recommended items, and thus, are suboptimal in helping users make informed decision on the recommendation platform. To tackle this problem, we propose a flexible model-agnostic method named MMI (Maximizing Mutual Information) framework to enhance the alignment between the generated natural language explanations and the predicted rating/important item features. Specifically, we propose to use mutual information (MI) as a measure for the alignment and train a neural MI estimator. Then, we treat a well-trained explanation generation model as the backbone model and further fine-tune it through reinforcement learning with guidance from the MI estimator, which rewards a generated explanation that is more aligned with the predicted rating or a pre-defined feature of the recommended item. Experiments on three datasets demonstrate that our MMI framework can boost different backbone models, enabling them to outperform existing baselines in terms of alignment with predicted ratings and item features. Additionally, user studies verify that MI-enhanced explanations indeed facilitate users' decisions and are favorable compared with other baselines due to their better alignment properties.
- Open Access
- 8 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 23-32
LARR: Large Language Model Aided Real-time Scene Recommendation with Semantic Understanding
- 8 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 23-32
Open AccessLARR: Large Language Model Aided Real-time Scene Recommendation with Semantic Understanding
- Open Access
- 24 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 6291-6300
Unified Dual-Intent Translation for Joint Modeling of Search and Recommendation
Recommendation systems, which assist users in discovering their preferred items among numerous options, have served billions of users across various online platforms. Intuitively, users' interactions with items are highly driven by their unchanging inherent intents (e.g., always preferring high-quality items) and changing demand intents (e.g., wanting a T-shirt in summer but a down jacket in winter). However, both types of intents are implicitly expressed in recommendation scenario, posing challenges in leveraging them for accurate intent-aware recommendations. Fortunately, in search scenario, often found alongside recommendation on the same online platform, users express their demand intents explicitly through their query words. Intuitively, in both scenarios, a user shares the same inherent intent and his/her interactions may be influenced by the same demand intent. It is therefore feasible to utilize the interaction data from both scenarios to reinforce the dual intents for joint intent-aware modeling. But the joint modeling should deal with two problems: (1) accurately modeling users' implicit demand intents in recommendation; (2) modeling the relation between the dual intents and the interactive items. To address these problems, we propose a novel model named Unified Dual-Intents Translation for joint modeling of Search and Recommendation (UDITSR). To accurately simulate users' demand intents in recommendation, we utilize real queries from search data as supervision information to guide its generation. To explicitly model the relation among the triplet , we propose a dual-intent translation propagation mechanism to learn the triplet in the same semantic space via embedding translations. Extensive experiments demonstrate that UDITSR outperforms SOTA baselines both in search and recommendation tasks. Moreover, our model has been deployed online on Meituan Waimai platform, leading to an average improvement in GMV (Gross Merchandise Value) of 1.46% and CTR(Click-Through Rate) of 0.77% over one month.
- 24 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 6291-6300
Open AccessUnified Dual-Intent Translation for Joint Modeling of Search and Recommendation
Recommendation systems, which assist users in discovering their preferred items among numerous options, have served billions of users across various online platforms. Intuitively, users' interactions with items are highly driven by their unchanging inherent intents (e.g., always preferring high-quality items) and changing demand intents (e.g., wanting a T-shirt in summer but a down jacket in winter). However, both types of intents are implicitly expressed in recommendation scenario, posing challenges in leveraging them for accurate intent-aware recommendations. Fortunately, in search scenario, often found alongside recommendation on the same online platform, users express their demand intents explicitly through their query words. Intuitively, in both scenarios, a user shares the same inherent intent and his/her interactions may be influenced by the same demand intent. It is therefore feasible to utilize the interaction data from both scenarios to reinforce the dual intents for joint intent-aware modeling. But the joint modeling should deal with two problems: (1) accurately modeling users' implicit demand intents in recommendation; (2) modeling the relation between the dual intents and the interactive items. To address these problems, we propose a novel model named Unified Dual-Intents Translation for joint modeling of Search and Recommendation (UDITSR). To accurately simulate users' demand intents in recommendation, we utilize real queries from search data as supervision information to guide its generation. To explicitly model the relation among the triplet , we propose a dual-intent translation propagation mechanism to learn the triplet in the same semantic space via embedding translations. Extensive experiments demonstrate that UDITSR outperforms SOTA baselines both in search and recommendation tasks. Moreover, our model has been deployed online on Meituan Waimai platform, leading to an average improvement in GMV (Gross Merchandise Value) of 1.46% and CTR(Click-Through Rate) of 0.77% over one month.
- Open Access
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 355-363
Context-based Fast Recommendation Strategy for Long User Behavior Sequence in Meituan Waimai
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 355-363
Open AccessContext-based Fast Recommendation Strategy for Long User Behavior Sequence in Meituan Waimai
- Open Access
- 14 September 2023
- Published by Association for Computing Machinery (ACM)
- p. 599-601
Heterogeneous Knowledge Fusion: A Novel Approach for Personalized Recommendation via LLM
- 14 September 2023
- Published by Association for Computing Machinery (ACM)
- p. 599-601
Open AccessHeterogeneous Knowledge Fusion: A Novel Approach for Personalized Recommendation via LLM