Yanwu Yang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yanwu Yang has a total of 30 co-authors. Here are the top co-authors by number of shared publications:
- Baozhu Feng
School of Management, Huazhong University of Science and Technology, Wuhan, China
2 shared publications - Bernard J. Jansen
Qatar Computing Research Institute, Hamad Bin Khalifa University
2 shared publications - Chunjie Zhang
Institute of Information Science, School of Computer Science and Technology, Beijing Jiaotong University, Beijing 100044, China
2 shared publications - Daniel Dajun Zeng
State Key Laboratory of Multimodal Artificial Intelligence Systems, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China
2 shared publications - Joni Salminen
School of Marketing and Communication, University of Vaasa
2 shared publications Panyu Zhai
2 shared publications
Latest Publications
- Open Access
- 20 October 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Knowledge Discovery From Data
- Vol. 20 (9) , 1-41
GraphLLM4CTR: Large Language Model-enhanced Graph Neural Networks for Click-through Rate Prediction in Recommender Systems
Click-through rate (CTR) prediction is a crucial task to estimate the probability that users will click on items in online platforms. Recently, researchers have incorporated semantics derived from large language models (LLMs) into representations of users and items in a graph neural network (GNN) framework to facilitate CTR prediction. However, existing hybrid GNN-LLM models predominantly treat GNNs and LLMs as two independent components and suffer from ineffective modality alignment between graph and textual representations. To address these issues, we propose LLMs-enhanced GNNs for CTR prediction (GraphLLM4CTR), where LLMs coordinate message passing in GNNs by infusing layer-wise semantics into message updates and graph representation learning. Specifically, we design a multi-view contrastive alignment (MVCA) method to transform graph and prompt-based embeddings into modality-aware representations and achieve modality alignment through intra-view and inter-view contrastive learning. Moreover, we propose an LLMs-based message coordinator (LBMC) that projects graph representations into an aligned latent space and incorporates LLMs with learnable adapters to regulate semantic-injected graph representation learning. Subsequently, we present a hybrid-expert mechanism with gating networks to adaptively aggregate representations for CTR prediction. Experiments conducted on three public datasets (i.e., MovieLens-1M, BookCrossing, and Amazon-Sports) demonstrate the superiority of GraphLLM4CTR over state-of-the-art baselines in terms of AUC and Logloss, and validate the adaptability of MVCA to topological properties and of LBMC to text qualities, and the dynamic balance of the hybrid-expert mechanism. Finally, we conduct an analysis of space and time complexity and computational efficiency of GraphLLM4CTR during training and inference stages, which indicates that GraphLLM4CTR can achieve the optimal AUC-latency tradeoff and linear scalability with the number of nodes and edges in constructed graphs.
- 20 October 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Knowledge Discovery From Data
- Vol. 20 (9) , 1-41
Open AccessGraphLLM4CTR: Large Language Model-enhanced Graph Neural Networks for Click-through Rate Prediction in Recommender Systems
Click-through rate (CTR) prediction is a crucial task to estimate the probability that users will click on items in online platforms. Recently, researchers have incorporated semantics derived from large language models (LLMs) into representations of users and items in a graph neural network (GNN) framework to facilitate CTR prediction. However, existing hybrid GNN-LLM models predominantly treat GNNs and LLMs as two independent components and suffer from ineffective modality alignment between graph and textual representations. To address these issues, we propose LLMs-enhanced GNNs for CTR prediction (GraphLLM4CTR), where LLMs coordinate message passing in GNNs by infusing layer-wise semantics into message updates and graph representation learning. Specifically, we design a multi-view contrastive alignment (MVCA) method to transform graph and prompt-based embeddings into modality-aware representations and achieve modality alignment through intra-view and inter-view contrastive learning. Moreover, we propose an LLMs-based message coordinator (LBMC) that projects graph representations into an aligned latent space and incorporates LLMs with learnable adapters to regulate semantic-injected graph representation learning. Subsequently, we present a hybrid-expert mechanism with gating networks to adaptively aggregate representations for CTR prediction. Experiments conducted on three public datasets (i.e., MovieLens-1M, BookCrossing, and Amazon-Sports) demonstrate the superiority of GraphLLM4CTR over state-of-the-art baselines in terms of AUC and Logloss, and validate the adaptability of MVCA to topological properties and of LBMC to text qualities, and the dynamic balance of the hybrid-expert mechanism. Finally, we conduct an analysis of space and time complexity and computational efficiency of GraphLLM4CTR during training and inference stages, which indicates that GraphLLM4CTR can achieve the optimal AUC-latency tradeoff and linear scalability with the number of nodes and edges in constructed graphs.
- 17 March 2026
- Published by Elsevier in Information Processing & Management
- Vol. 63 (6)
MLAFormer: Multi-scale transformer with local convolutional auto-correlation and pre-training for time series forecasting
- 17 March 2026
- Published by Elsevier in Information Processing & Management
- Vol. 63 (6)
MLAFormer: Multi-scale transformer with local convolutional auto-correlation and pre-training for time series forecasting
- Open Access
- 14 November 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
- Vol. 44 (1) , 1-37
Periodic Graph Neural Networks for Click-Through Rate Prediction in Online Advertising
CTR prediction serves as a valuable function to provide indications about the effectiveness of advertising campaigns. Numerous models have been developed to learn dynamic representations and sophisticated feature interactions for CTR prediction. We observe that users’ behavioral sequences have periodic patterns, which is a crucial factor for capturing the temporal dependency in highly dynamic environments such as online advertising. Unfortunately, existing work ignores periodic patterns in CTR prediction, and thus incurs the low model performance. This article proposes a periodic CTR prediction model in the GNNs modeling framework (PGNN), that combines periodic graph representations and feature graph representations. The former learns periodic graph representations of users’ and ads’ sequences with multi-scale periodic patterns and the high-order collaborative information across sequences on a dynamic graph of user-ad interactions, and the latter is designed to learn sophisticated feature interactions by incorporating the principle of field-aware feature interaction into an interpolable graph convolutional attention mechanism on a feature graph. Experiments conducted on three public datasets (i.e., Movielens-1M, Criteo-attribution, and Alimama) demonstrate the superiority of PGNN. PGNN outperforms the strongest baseline by 0.01–0.02 in terms of AUC and Logloss. Meanwhile, the effectiveness of periodic graph representation learning is also verified in this research.
- 14 November 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
- Vol. 44 (1) , 1-37
Open AccessPeriodic Graph Neural Networks for Click-Through Rate Prediction in Online Advertising
CTR prediction serves as a valuable function to provide indications about the effectiveness of advertising campaigns. Numerous models have been developed to learn dynamic representations and sophisticated feature interactions for CTR prediction. We observe that users’ behavioral sequences have periodic patterns, which is a crucial factor for capturing the temporal dependency in highly dynamic environments such as online advertising. Unfortunately, existing work ignores periodic patterns in CTR prediction, and thus incurs the low model performance. This article proposes a periodic CTR prediction model in the GNNs modeling framework (PGNN), that combines periodic graph representations and feature graph representations. The former learns periodic graph representations of users’ and ads’ sequences with multi-scale periodic patterns and the high-order collaborative information across sequences on a dynamic graph of user-ad interactions, and the latter is designed to learn sophisticated feature interactions by incorporating the principle of field-aware feature interaction into an interpolable graph convolutional attention mechanism on a feature graph. Experiments conducted on three public datasets (i.e., Movielens-1M, Criteo-attribution, and Alimama) demonstrate the superiority of PGNN. PGNN outperforms the strongest baseline by 0.01–0.02 in terms of AUC and Logloss. Meanwhile, the effectiveness of periodic graph representation learning is also verified in this research.
- 7 November 2025
- Published by Springer Nature in Electronic Commerce Research
- Vol. 26 (5) , 1-50
Toward a benchmark for CTR prediction in online advertising: datasets, evaluation protocols and perspectives
- 7 November 2025
- Published by Springer Nature in Electronic Commerce Research
- Vol. 26 (5) , 1-50
Toward a benchmark for CTR prediction in online advertising: datasets, evaluation protocols and perspectives
- 31 October 2025
- Published by Elsevier in Information Processing & Management
- Vol. 63 (2)
Conversion rate prediction in online advertising: modeling techniques, performance evaluation and future directions
- 31 October 2025
- Published by Elsevier in Information Processing & Management
- Vol. 63 (2)
Conversion rate prediction in online advertising: modeling techniques, performance evaluation and future directions
- 18 May 2025
- Published by Springer Nature in Electronic Commerce Research
- Vol. 26 (2) , 2211-2240
Dynamic heterogeneous graph convolutional networks for click-through rate prediction in recommender systems
- 18 May 2025
- Published by Springer Nature in Electronic Commerce Research
- Vol. 26 (2) , 2211-2240
Dynamic heterogeneous graph convolutional networks for click-through rate prediction in recommender systems
- Open Access
- 10 September 2024
- Published by Association for Computing Machinery (ACM)
- p. 323-329
Engagement Patterns in TikTok: An Analysis of Short Video Ads
Short videos have become one of the dominant formats of social media content. Such short videos are often boosted by advertising to increase their reach and visibility. We investigate user churn (i.e., drop or loss of consumers during the video viewing) of short video ads on TikTok. Data from the TikTok Ads account of an e-commerce company was analyzed using logistic regression to examine how gender, age groups, and video duration affect users’ video viewing behaviors. The findings indicate that (1) most customers leave the ads within the first quarter of the video; (2) males and females exhibit similar viewing behaviors, but (3) age groups vary by their viewing behaviors. Somewhat surprisingly, we discovered that (4) there is no significant correlation between user churn and ads performance metrics (i.e., cost-per-mille, cost-per-click, and click-through rate).
- 10 September 2024
- Published by Association for Computing Machinery (ACM)
- p. 323-329
Open AccessEngagement Patterns in TikTok: An Analysis of Short Video Ads
Short videos have become one of the dominant formats of social media content. Such short videos are often boosted by advertising to increase their reach and visibility. We investigate user churn (i.e., drop or loss of consumers during the video viewing) of short video ads on TikTok. Data from the TikTok Ads account of an e-commerce company was analyzed using logistic regression to examine how gender, age groups, and video duration affect users’ video viewing behaviors. The findings indicate that (1) most customers leave the ads within the first quarter of the video; (2) males and females exhibit similar viewing behaviors, but (3) age groups vary by their viewing behaviors. Somewhat surprisingly, we discovered that (4) there is no significant correlation between user churn and ads performance metrics (i.e., cost-per-mille, cost-per-click, and click-through rate).
- 17 February 2024
- Published by Wiley in Psychology & Marketing
- Vol. 41 (6) , 1289-1317
How do username and avatar affect people's engagement with native advertising on social media: From the self‐disclosure perspective
The effect of self‐profiling on people's responses toward commercial information (e.g., advertisements) delivered on social media remains unexplored. This study proposed a conceptual model examining the path from the authenticity of username and avatar and its antecedents to native advertising engagement from the perspective of self‐disclosure. Four subsequent studies were conducted to test our research model. The results indicated that the authenticity of username and avatar is a crucial determinant of engagement with advertising information. Moreover, social media involvement moderates the relationship between the authenticity of username and avatar and advertising engagement. Social media engagement partially mediates the relationship between the authenticity of username and avatar and advertising engagement. Regarding the antecedents of the authenticity of username and avatar, self‐presentation, relationship building, privacy concerns, privacy risks and trust belief are significant, while relationship maintenance is insignificant. This study contributes to our understanding of self‐profiling authenticity and its effect on people's engagement by employing the theory of self‐disclosure in the context of social media and provides critical managerial insights for social media platforms and advertisers.
- 17 February 2024
- Published by Wiley in Psychology & Marketing
- Vol. 41 (6) , 1289-1317
How do username and avatar affect people's engagement with native advertising on social media: From the self‐disclosure perspective
The effect of self‐profiling on people's responses toward commercial information (e.g., advertisements) delivered on social media remains unexplored. This study proposed a conceptual model examining the path from the authenticity of username and avatar and its antecedents to native advertising engagement from the perspective of self‐disclosure. Four subsequent studies were conducted to test our research model. The results indicated that the authenticity of username and avatar is a crucial determinant of engagement with advertising information. Moreover, social media involvement moderates the relationship between the authenticity of username and avatar and advertising engagement. Social media engagement partially mediates the relationship between the authenticity of username and avatar and advertising engagement. Regarding the antecedents of the authenticity of username and avatar, self‐presentation, relationship building, privacy concerns, privacy risks and trust belief are significant, while relationship maintenance is insignificant. This study contributes to our understanding of self‐profiling authenticity and its effect on people's engagement by employing the theory of self‐disclosure in the context of social media and provides critical managerial insights for social media platforms and advertisers.
- 10 November 2022
- Published by Elsevier in Information Processing & Management
- Vol. 60 (1)
Causality-based CTR prediction using graph neural networks
- 10 November 2022
- Published by Elsevier in Information Processing & Management
- Vol. 60 (1)
Causality-based CTR prediction using graph neural networks
- 8 November 2022
- Published by Elsevier in Information Processing & Management
- Vol. 60 (1)
Anchor-free temporal action localization via Progressive Boundary-aware Boosting
- 8 November 2022
- Published by Elsevier in Information Processing & Management
- Vol. 60 (1)
Anchor-free temporal action localization via Progressive Boundary-aware Boosting