Xiaolu Zhang
Xiaolu Zhang has a total of 66 co-authors. Here are the top co-authors by number of shared publications:
- Jun Zhou
Ant Group, Hangzhou, Zhejiang, China
21 shared publications - Zhaoxin Huan
Ant Group CO Ltd, Hangzhou, China
10 shared publications - Yong He
Ant Group, Hangzhou, China
8 shared publications - Chilin Fu
Ant Group CO Ltd, Hangzhou, China
7 shared publications - Chenliang Li
Wuhan University
6 shared publications - Jun Hu
Ant Group CO Ltd, Hangzhou, China
6 shared publications
Latest Publications
- Open Access
- 17 April 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
- Vol. 44 (4) , 1
Corrigendum: One Model for All: Large Language Models Are Domain-Agnostic Recommendation Systems
This is a corrigendum for the article “One Model for All: Large Language Models are Domain-Agnostic Recommendation Systems” published in ACM Trans. Inf. Syst . 43, 5, Article 118 (July 2025), 27 pages.
- 17 April 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
- Vol. 44 (4) , 1
Open AccessCorrigendum: One Model for All: Large Language Models Are Domain-Agnostic Recommendation Systems
This is a corrigendum for the article “One Model for All: Large Language Models are Domain-Agnostic Recommendation Systems” published in ACM Trans. Inf. Syst . 43, 5, Article 118 (July 2025), 27 pages.
- Top Cited PapersOpen Access
- 10 July 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
- Vol. 43 (5) , 1-27
One Model for All: Large Language Models Are Domain-Agnostic Recommendation Systems
Sequential recommendation systems aim to predict users’ next likely interaction based on their history. However, these systems face data sparsity and cold-start problems. Utilizing data from other domains, known as multi-domain methods, is useful for alleviating these problems. However, traditional multi-domain methods rely on meaningless ID-based item representation, which makes it difficult to align items with similar meanings from different domains, yielding sup-optimal knowledge transfer. This article introduces LLM-Rec , a framework that utilizes pre-trained Large Language Models (LLMs) for domain-agnostic recommendation. Specifically, we mix user’s behaviors from multiple domains and concatenate item titles into a sentence, then use LLMs for generating user and item representations. By mixing behaviors across different domains, we can exploit the knowledge encoded in LLMs to bridge the semantic across over multi-domain behaviors, thus obtaining semantically rich representations and improving performance in all domains. Furthermore, we explore the underlying reasons why LLMs are effective and investigate whether LLMs can understand the semantic correlations as the recommendation model, and if advanced techniques like scaling laws in NLP also work in recommendations. We conduct extensive experiments with LLMs ranging from 40 M to 6.7 B to answer the above questions and to verify the effectiveness of LLM-Rec in multi-domain recommendation. The source code is available at https://github.com/WHUIR/LLMRec .
- 10 July 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
- Vol. 43 (5) , 1-27
Top Cited PapersOpen AccessOne Model for All: Large Language Models Are Domain-Agnostic Recommendation Systems
Sequential recommendation systems aim to predict users’ next likely interaction based on their history. However, these systems face data sparsity and cold-start problems. Utilizing data from other domains, known as multi-domain methods, is useful for alleviating these problems. However, traditional multi-domain methods rely on meaningless ID-based item representation, which makes it difficult to align items with similar meanings from different domains, yielding sup-optimal knowledge transfer. This article introduces LLM-Rec , a framework that utilizes pre-trained Large Language Models (LLMs) for domain-agnostic recommendation. Specifically, we mix user’s behaviors from multiple domains and concatenate item titles into a sentence, then use LLMs for generating user and item representations. By mixing behaviors across different domains, we can exploit the knowledge encoded in LLMs to bridge the semantic across over multi-domain behaviors, thus obtaining semantically rich representations and improving performance in all domains. Furthermore, we explore the underlying reasons why LLMs are effective and investigate whether LLMs can understand the semantic correlations as the recommendation model, and if advanced techniques like scaling laws in NLP also work in recommendations. We conduct extensive experiments with LLMs ranging from 40 M to 6.7 B to answer the above questions and to verify the effectiveness of LLM-Rec in multi-domain recommendation. The source code is available at https://github.com/WHUIR/LLMRec .
- Open Access
- 9 May 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
- Vol. 43 (4) , 1-29
HORAE: Temporal Multi-Interest Pre-training for Sequential Recommendation
The data sparsity problem has been a long-standing obstacle towards achieving better recommendation performance since it is miserable to estimate the user’s interests from limited historical behaviors. The pre-training paradigm, i.e., learning universal knowledge across a wide spectrum of domains, has increasingly become a new de-facto practice in many fields, especially for adaption to new domains. The merit of this superior generalizability renders it a natural choice to tackle the data sparsity problem for various recommendation scenarios. Hence, several efforts mainly follow masked language modeling or simple data augmentation via contrastive learning to build a pre-trained recommendation model. Our recent work (namely Miracle ) suggests that the common treatment utilizing the masked language modeling is not sufficient for pre-training a recommender system, since a user’s intent could be more complex than predicting the next word or item. The encouraging results demonstrate that the multi-interest modeling could significantly push the frontier of recommender system pre-training. Nevertheless, how to accommodate the temporal dynamics of the user interests seems to be underexplored under both single vector representation and multi-interest schemes. In this article, we aim to incorporate sophisticated temporal information modeling with the current advance in this line. More specifically, we extend Miracle by further considering relative position information and two kinds of relative time interval information jointly when performing multi-interest learning. Then, a sequential process for interest refinement is proposed to learn the subtle nuances of how interests change and shift along the timeline, leading to a more precise representation of user interests. Our extensive experiments on multiple real-world datasets validate the effectiveness of the proposed solution, demonstrating a significant improvement over current state-of-the-art models on these benchmarks. The code is available at https://github.com/WHUIR/Horae .
- 9 May 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
- Vol. 43 (4) , 1-29
Open AccessHORAE: Temporal Multi-Interest Pre-training for Sequential Recommendation
The data sparsity problem has been a long-standing obstacle towards achieving better recommendation performance since it is miserable to estimate the user’s interests from limited historical behaviors. The pre-training paradigm, i.e., learning universal knowledge across a wide spectrum of domains, has increasingly become a new de-facto practice in many fields, especially for adaption to new domains. The merit of this superior generalizability renders it a natural choice to tackle the data sparsity problem for various recommendation scenarios. Hence, several efforts mainly follow masked language modeling or simple data augmentation via contrastive learning to build a pre-trained recommendation model. Our recent work (namely Miracle ) suggests that the common treatment utilizing the masked language modeling is not sufficient for pre-training a recommender system, since a user’s intent could be more complex than predicting the next word or item. The encouraging results demonstrate that the multi-interest modeling could significantly push the frontier of recommender system pre-training. Nevertheless, how to accommodate the temporal dynamics of the user interests seems to be underexplored under both single vector representation and multi-interest schemes. In this article, we aim to incorporate sophisticated temporal information modeling with the current advance in this line. More specifically, we extend Miracle by further considering relative position information and two kinds of relative time interval information jointly when performing multi-interest learning. Then, a sequential process for interest refinement is proposed to learn the subtle nuances of how interests change and shift along the timeline, leading to a more precise representation of user interests. Our extensive experiments on multiple real-world datasets validate the effectiveness of the proposed solution, demonstrating a significant improvement over current state-of-the-art models on these benchmarks. The code is available at https://github.com/WHUIR/Horae .
- Open Access
- 8 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 325-333
SurvFormer: A Transformer Based Framework for Survival Analysis in Insurance Underwriting
- 8 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 325-333
Open AccessSurvFormer: A Transformer Based Framework for Survival Analysis in Insurance Underwriting
- Open Access
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 3063-3072
MMLRec: A Unified Multi-Task and Multi-Scenario Learning Benchmark for Recommendation
In recent years, there has been a trend in the field of recommender systems towards multi-task modeling and multi-scenario modeling. The aim is to enhance the performance of various tasks and scenarios by jointly training on multiple tasks or scenarios to learn common patterns and features. Joint modeling of tasks and scenarios has also received widespread attention recently. However, despite the rich proposals of methods for Multi-Task Learning (MTL), Multi-Scenario Learning (MSL), and Multi-Task-Multi-Scenario Learning (MTMSL) in recent years, there still lacks a comprehensive benchmark to evaluate these methods. Previous studies often employed different datasets, data processing techniques, data partitioning strategies, and hyperparameter settings, making replication of existing research and fair comparison of experimental results challenging. To address this challenge, we introduce MMLRec, the first unified comprehensive benchmark for evaluating MTL, MSL and MTMSL, featuring consistent dataset processing and identical parameter settings. This benchmark implements a range of MTL, MSL, and MTMSL algorithms, and evaluates them on multiple commonly used recommender systems datasets. Through fair comparative experiments, we find that some structurally simplistic recommendation algorithms are underestimated, as they can achieve comparable results to more complex algorithms while maintaining lower complexity. Furthermore, our experimental analysis indicates that more complex methods exhibit better robustness when there are significant differences between tasks or scenarios. By providing a unified framework (MMLRec), our goal is to promote rapid evaluation and inspire innovative research in this continuously evolving field. We hope that our open-source benchmark can facilitate swift, equitable evaluations, while also fostering further breakthrough research in the domains of MTL, MSL, and MTMSL.
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 3063-3072
Open AccessMMLRec: A Unified Multi-Task and Multi-Scenario Learning Benchmark for Recommendation
In recent years, there has been a trend in the field of recommender systems towards multi-task modeling and multi-scenario modeling. The aim is to enhance the performance of various tasks and scenarios by jointly training on multiple tasks or scenarios to learn common patterns and features. Joint modeling of tasks and scenarios has also received widespread attention recently. However, despite the rich proposals of methods for Multi-Task Learning (MTL), Multi-Scenario Learning (MSL), and Multi-Task-Multi-Scenario Learning (MTMSL) in recent years, there still lacks a comprehensive benchmark to evaluate these methods. Previous studies often employed different datasets, data processing techniques, data partitioning strategies, and hyperparameter settings, making replication of existing research and fair comparison of experimental results challenging. To address this challenge, we introduce MMLRec, the first unified comprehensive benchmark for evaluating MTL, MSL and MTMSL, featuring consistent dataset processing and identical parameter settings. This benchmark implements a range of MTL, MSL, and MTMSL algorithms, and evaluates them on multiple commonly used recommender systems datasets. Through fair comparative experiments, we find that some structurally simplistic recommendation algorithms are underestimated, as they can achieve comparable results to more complex algorithms while maintaining lower complexity. Furthermore, our experimental analysis indicates that more complex methods exhibit better robustness when there are significant differences between tasks or scenarios. By providing a unified framework (MMLRec), our goal is to promote rapid evaluation and inspire innovative research in this continuously evolving field. We hope that our open-source benchmark can facilitate swift, equitable evaluations, while also fostering further breakthrough research in the domains of MTL, MSL, and MTMSL.
- Open Access
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 2240-2249
TEXT CAN BE FAIR: Mitigating Popularity Bias with PLMs by Learning Relative Preference
Recently, the item textual information has been exploited with pre-trained language models (PLMs) to enrich the representations of tail items. The underlying idea is to align the hot items and tail items in terms of the external semantic knowledge covered by the PLM. However, it is non-trivial to eliminate the popularity bias by exploiting the textual semantics. One major obstacle is that the model supervision still counts on the sparse yet binary user behaviors. In the preliminary investigation, we discover that text-based recommendations also suffer from the popularity bias. To this end, we propose a novel self-distillation framework based on a pre-trained language model, named Staple. The proposed Staple consists of two main components: ranker model and recommender model, which are both instantiated as a PLM towards exploiting the item textual semantics. Motivated by the recent success of reinforcement learning with human feedback (RLHF), the proposed Staple aims to recover the relative preference by learning a fair ranker model that can successfully distinguish the preference levels for uninteracted items. Specifically, analogous to the training of large language models (LLMs), we introduce a pre-training and a fair supervised fine-tuning with a decoupled layer to build the ranker model. Then, similar to RLHF for LLM training, we utilize the relative preference information estimated by the ranker over candidate items to complement the learning of the recommender model. We show that this RLHF process can be reformed as an efficient distillation learning process. We conduct extensive experiments on three real-world datasets. In addition to the performance metrics, we employ two additional metrics to measure fairness and debiased performance. The experiments show that our method can significantly improve the item exposure fairness of recommendation and mitigate popularity bias, while also improving the recommendation performance. The source code is available at https://github.com/WHUIR/STAPLE.
- 21 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 2240-2249
Open AccessTEXT CAN BE FAIR: Mitigating Popularity Bias with PLMs by Learning Relative Preference
Recently, the item textual information has been exploited with pre-trained language models (PLMs) to enrich the representations of tail items. The underlying idea is to align the hot items and tail items in terms of the external semantic knowledge covered by the PLM. However, it is non-trivial to eliminate the popularity bias by exploiting the textual semantics. One major obstacle is that the model supervision still counts on the sparse yet binary user behaviors. In the preliminary investigation, we discover that text-based recommendations also suffer from the popularity bias. To this end, we propose a novel self-distillation framework based on a pre-trained language model, named Staple. The proposed Staple consists of two main components: ranker model and recommender model, which are both instantiated as a PLM towards exploiting the item textual semantics. Motivated by the recent success of reinforcement learning with human feedback (RLHF), the proposed Staple aims to recover the relative preference by learning a fair ranker model that can successfully distinguish the preference levels for uninteracted items. Specifically, analogous to the training of large language models (LLMs), we introduce a pre-training and a fair supervised fine-tuning with a decoupled layer to build the ranker model. Then, similar to RLHF for LLM training, we utilize the relative preference information estimated by the ranker over candidate items to complement the learning of the recommender model. We show that this RLHF process can be reformed as an efficient distillation learning process. We conduct extensive experiments on three real-world datasets. In addition to the performance metrics, we employ two additional metrics to measure fairness and debiased performance. The experiments show that our method can significantly improve the item exposure fairness of recommendation and mitigate popularity bias, while also improving the recommendation performance. The source code is available at https://github.com/WHUIR/STAPLE.
- Open Access
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 1232-1241
Exploring Multi-Scenario Multi-Modal CTR Prediction with a Large Scale Dataset
Click-through rate (CTR) prediction plays a crucial role in recommendation systems, with significant impact on user experience and platform revenue generation. Despite the various public CTR datasets available due to increasing interest from both academia and industry, these datasets have limitations. They cover a limited range of scenarios and predominantly focus on ID-based features, neglecting the vital role of multi-modal features for effective multi-scenario CTR prediction. Moreover, their scale is modest compared to real-world industrial datasets, hindering robust and comprehensive evaluation of complex models. To address these challenges, we introduce a large-scale Multi-Scenario Multi-Modal CTR dataset named AntM2 C, built from real industrial data from Alipay. This dataset offers an impressive breadth and depth of information, covering CTR data from four diverse business scenarios, including advertisements, consumer coupons, mini-programs, and videos. Unlike existing datasets, AntM2 C provides not only ID-based features but also five textual features and one image feature for both users and items, supporting more delicate multi-modal CTR prediction. AntM2 C is also substantially larger than existing datasets, comprising 100 million CTR data. This scale allows for robust and comprehensive evaluation and comparison of CTR prediction models. We employ AntM2 C to construct several typical CTR tasks, including multi-scenario modeling, item and user cold-start modeling, and multi-modal modeling. Initial experiments and comparisons with baseline methods have shown that AntM2 C presents both new challenges and opportunities for CTR models, with the potential to significantly advance CTR research. The AntM2 C dataset is available at https://www.atecup.cn/OfficalDataSet.
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 1232-1241
Open AccessExploring Multi-Scenario Multi-Modal CTR Prediction with a Large Scale Dataset
Click-through rate (CTR) prediction plays a crucial role in recommendation systems, with significant impact on user experience and platform revenue generation. Despite the various public CTR datasets available due to increasing interest from both academia and industry, these datasets have limitations. They cover a limited range of scenarios and predominantly focus on ID-based features, neglecting the vital role of multi-modal features for effective multi-scenario CTR prediction. Moreover, their scale is modest compared to real-world industrial datasets, hindering robust and comprehensive evaluation of complex models. To address these challenges, we introduce a large-scale Multi-Scenario Multi-Modal CTR dataset named AntM2 C, built from real industrial data from Alipay. This dataset offers an impressive breadth and depth of information, covering CTR data from four diverse business scenarios, including advertisements, consumer coupons, mini-programs, and videos. Unlike existing datasets, AntM2 C provides not only ID-based features but also five textual features and one image feature for both users and items, supporting more delicate multi-modal CTR prediction. AntM2 C is also substantially larger than existing datasets, comprising 100 million CTR data. This scale allows for robust and comprehensive evaluation and comparison of CTR prediction models. We employ AntM2 C to construct several typical CTR tasks, including multi-scenario modeling, item and user cold-start modeling, and multi-modal modeling. Initial experiments and comparisons with baseline methods have shown that AntM2 C presents both new challenges and opportunities for CTR models, with the potential to significantly advance CTR research. The AntM2 C dataset is available at https://www.atecup.cn/OfficalDataSet.
- Open Access
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 2311-2315
Breaking the Length Barrier: LLM-Enhanced CTR Prediction in Long Textual User Behaviors
With the rise of large language models (LLMs), recent works have leveraged LLMs to improve the performance of click-through rate (CTR) prediction. However, we argue that a critical obstacle remains in deploying LLMs for practical use: the efficiency of LLMs when processing long textual user behaviors. As user sequences grow longer, the current efficiency of LLMs is inadequate for training on billions of users and items. To break through the efficiency barrier of LLMs, we propose Behavior Aggregated Hierarchical Encoding (BAHE) to enhance the efficiency of LLM-based CTR modeling. Specifically, BAHE proposes a novel hierarchical architecture that decouples the encoding of user behaviors from inter-behavior interactions. Firstly, to prevent computational redundancy from repeated encoding of identical user behaviors, BAHE employs the LLM's pre-trained shallow layers to extract embeddings of the most granular, atomic user behaviors from extensive user sequences and stores them in the offline database. Subsequently, the deeper, trainable layers of the LLM facilitate intricate inter-behavior interactions, thereby generating comprehensive user embeddings. This separation allows the learning of high-level user representations to be independent of low-level behavior encoding, significantly reducing computational complexity. Finally, these refined user embeddings, in conjunction with correspondingly processed item embeddings, are incorporated into the CTR model to compute the CTR scores. Extensive experimental results show that BAHE reduces training time and memory by five times for CTR models using LLMs, especially with longer user sequences. BAHE has been deployed in a real-world system, allowing for daily updates of 50 million CTR data on 8 A100 GPUs, making LLMs practical for industrial CTR prediction.
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 2311-2315
Open AccessBreaking the Length Barrier: LLM-Enhanced CTR Prediction in Long Textual User Behaviors
With the rise of large language models (LLMs), recent works have leveraged LLMs to improve the performance of click-through rate (CTR) prediction. However, we argue that a critical obstacle remains in deploying LLMs for practical use: the efficiency of LLMs when processing long textual user behaviors. As user sequences grow longer, the current efficiency of LLMs is inadequate for training on billions of users and items. To break through the efficiency barrier of LLMs, we propose Behavior Aggregated Hierarchical Encoding (BAHE) to enhance the efficiency of LLM-based CTR modeling. Specifically, BAHE proposes a novel hierarchical architecture that decouples the encoding of user behaviors from inter-behavior interactions. Firstly, to prevent computational redundancy from repeated encoding of identical user behaviors, BAHE employs the LLM's pre-trained shallow layers to extract embeddings of the most granular, atomic user behaviors from extensive user sequences and stores them in the offline database. Subsequently, the deeper, trainable layers of the LLM facilitate intricate inter-behavior interactions, thereby generating comprehensive user embeddings. This separation allows the learning of high-level user representations to be independent of low-level behavior encoding, significantly reducing computational complexity. Finally, these refined user embeddings, in conjunction with correspondingly processed item embeddings, are incorporated into the CTR model to compute the CTR scores. Extensive experimental results show that BAHE reduces training time and memory by five times for CTR models using LLMs, especially with longer user sequences. BAHE has been deployed in a real-world system, allowing for daily updates of 50 million CTR data on 8 A100 GPUs, making LLMs practical for industrial CTR prediction.
- Open Access
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 103-111
Enhancing Sequential Recommendation via LLM-based Semantic Embedding Learning
Sequential recommendation systems (SRS) are crucial in various applications as they enable users to discover relevant items based on their past interactions. Recent advancements involving large language models (LLMs) have shown significant promise in addressing intricate recommendation challenges. However, these efforts exhibit certain limitations. Specifically, directly extracting representations from an LLM based on items' textual features and feeding them into a sequential model hold no guarantee that the semantic information of texts could be preserved in these representations. Additionally, concatenating textual descriptions of all items in an item sequence into a long text and feeding it into an LLM for recommendation results in lengthy token sequences, which largely diminishes the practical efficiency. In this paper, we introduce SAID, a framework that utilizes LLMs to explicitly learn Semantically Aligned item ID embeddings based on texts. For each item, SAID employs a projector module to transform an item ID into an embedding vector, which will be fed into an LLM to elicit the exact descriptive text tokens accompanied by the item. The item embeddings are forced to preserve fine-grained semantic information of textual descriptions. Further, the learned embeddings can be integrated with lightweight downstream sequential models for practical recommendations. In this way, SAID circumvents lengthy token sequences in previous works, reducing resources required in industrial scenarios and also achieving superior recommendation performance. Experiments on six public datasets demonstrate that SAID outperforms baselines by about 5% to 15% in terms of NDCG@10. Moreover, SAID has been deployed in Alipay's online advertising platform, achieving a 3.07% relative improvement of cost per mille (CPM) over baselines, with an online response time of under 20 milliseconds.
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 103-111
Open AccessEnhancing Sequential Recommendation via LLM-based Semantic Embedding Learning
Sequential recommendation systems (SRS) are crucial in various applications as they enable users to discover relevant items based on their past interactions. Recent advancements involving large language models (LLMs) have shown significant promise in addressing intricate recommendation challenges. However, these efforts exhibit certain limitations. Specifically, directly extracting representations from an LLM based on items' textual features and feeding them into a sequential model hold no guarantee that the semantic information of texts could be preserved in these representations. Additionally, concatenating textual descriptions of all items in an item sequence into a long text and feeding it into an LLM for recommendation results in lengthy token sequences, which largely diminishes the practical efficiency. In this paper, we introduce SAID, a framework that utilizes LLMs to explicitly learn Semantically Aligned item ID embeddings based on texts. For each item, SAID employs a projector module to transform an item ID into an embedding vector, which will be fed into an LLM to elicit the exact descriptive text tokens accompanied by the item. The item embeddings are forced to preserve fine-grained semantic information of textual descriptions. Further, the learned embeddings can be integrated with lightweight downstream sequential models for practical recommendations. In this way, SAID circumvents lengthy token sequences in previous works, reducing resources required in industrial scenarios and also achieving superior recommendation performance. Experiments on six public datasets demonstrate that SAID outperforms baselines by about 5% to 15% in terms of NDCG@10. Moreover, SAID has been deployed in Alipay's online advertising platform, achieving a 3.07% relative improvement of cost per mille (CPM) over baselines, with an online response time of under 20 milliseconds.
- Open Access
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 112-120
Aligned Side Information Fusion Method for Sequential Recommendation
Combining contextual information (i.e., side information) of items beyond IDs has become an important way to improve the performance in recommender systems. Existing self-attention-based side information fusion methods can be categorized into early, late, and hybrid fusion. In practice, naive early fusion may interfere with the representation of IDs, resulting in negative effects, while late fusion misses effective interactions between IDs and side information. Some hybrid methods have been proposed to address these issues, but they only utilize side information in calculating attention scores, which may lead to information loss. To harness the full potential of side information without noisy interference, we propose an Aligned Side Information Fusion (ASIF) method for sequential recommendation, consisting of two parts: Fused Attention with Untied Positions and Representation Alignment. Specifically, we first decouple the positions to exclude the noisy interference in the attention scores. Secondly, we adopt the contrastive objective to maintain the semantic consistency between IDs and side information and then employ orthogonal decomposition to extract the homogeneous parts. By aligning the representations and fusing them together, ASIF makes full use of the side information without interfering with IDs. Offline experimental results on four datasets demonstrate the superiority of ASIF. Additionally, we successfully deployed the model in Alipay's advertising system and achieved 1.09% and 1.86% improvements on clicks and Cost Per Mille (CPM).
- 13 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 112-120
Open AccessAligned Side Information Fusion Method for Sequential Recommendation
Combining contextual information (i.e., side information) of items beyond IDs has become an important way to improve the performance in recommender systems. Existing self-attention-based side information fusion methods can be categorized into early, late, and hybrid fusion. In practice, naive early fusion may interfere with the representation of IDs, resulting in negative effects, while late fusion misses effective interactions between IDs and side information. Some hybrid methods have been proposed to address these issues, but they only utilize side information in calculating attention scores, which may lead to information loss. To harness the full potential of side information without noisy interference, we propose an Aligned Side Information Fusion (ASIF) method for sequential recommendation, consisting of two parts: Fused Attention with Untied Positions and Representation Alignment. Specifically, we first decouple the positions to exclude the noisy interference in the attention scores. Secondly, we adopt the contrastive objective to maintain the semantic consistency between IDs and side information and then employ orthogonal decomposition to extract the homogeneous parts. By aligning the representations and fusing them together, ASIF makes full use of the side information without interfering with IDs. Offline experimental results on four datasets demonstrate the superiority of ASIF. Additionally, we successfully deployed the model in Alipay's advertising system and achieved 1.09% and 1.86% improvements on clicks and Cost Per Mille (CPM).