Vincent W. Zheng
Vincent W. Zheng has a total of 62 co-authors. Here are the top co-authors by number of shared publications:
- Kevin Chen-Chuan Chang
University of Illinois at Urbana-Champaign USA
7 shared publications - Huan Zhao
Paradigm Inc., Beijing, China
3 shared publications - Wenyi Xiao
Huawei Technologies Co Ltd
3 shared publications - Yangqiu Song
Department of Computer Science and Engineering, Hong Kong University of Science and Technology, Hong Kong, Hong Kong
3 shared publications - Fanwei Zhu
Hangzhou City University, Hangzhou, China
2 shared publications - Hongyun Cai
Advanced Digital Sciences Center Singapore
2 shared publications
Latest Publications
- 25 July 2025
- Published by Springer Nature
- p. 57-87
Federated Learning for Enterprise AI
- Open Access
- 20 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 2271-2281
Efficient Multi-Expert Tabular Language Model for Banking
- 20 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 2271-2281
Open AccessEfficient Multi-Expert Tabular Language Model for Banking
- 18 December 2024
- Published by Springer Nature in Data Mining and Knowledge Discovery
- Vol. 39 (1) , 1-37
Meta-path based proximity learning in heterogeneous information networks
- 18 December 2024
- Published by Springer Nature in Data Mining and Knowledge Discovery
- Vol. 39 (1) , 1-37
Meta-path based proximity learning in heterogeneous information networks
- 1 August 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Learning Technologies
- Vol. 15 (4) , 467-480
CMKT: Concept Map Driven Knowledge Tracing
In this article, we advocate for and propose a novel concept map driven knowledge tracing (CMKT) model, which utilizes educational concept map for learner modeling. This article particularly addresses the issue of learner data sparseness caused by the unwillingness to practice and irregular learning behaviors on the learner side. CMKT considers the concept map as a new information source and explicitly exploits its inherent information to help the estimation of the learner's knowledge state. Specifically, the pairwise educational relations in the concept map are formulated as the ordering pairs and are used as mathematical constraints for model construction. The topology information in the concept map is extracted and used as the model input by employing the network embedding techniques. Integrating both educational relation information and topology information in the concept map, CMKT adopts the recurrent neural network to perform knowledge tracing tasks. Comprehensive evaluations conducted on five public educational datasets of four different subjects (more than 8000 learners and their 300 000 records) demonstrate the promise and effectiveness of CMKT: The average area under ROC curve (AUC) and overall prediction accuracy (ACC) achieve 0.82 and 0.75, respectively, and CMKT outperforms all the baselines by at least 12.2% and 9.2% in terms of AUC and ACC.
- 1 August 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Learning Technologies
- Vol. 15 (4) , 467-480
CMKT: Concept Map Driven Knowledge Tracing
In this article, we advocate for and propose a novel concept map driven knowledge tracing (CMKT) model, which utilizes educational concept map for learner modeling. This article particularly addresses the issue of learner data sparseness caused by the unwillingness to practice and irregular learning behaviors on the learner side. CMKT considers the concept map as a new information source and explicitly exploits its inherent information to help the estimation of the learner's knowledge state. Specifically, the pairwise educational relations in the concept map are formulated as the ordering pairs and are used as mathematical constraints for model construction. The topology information in the concept map is extracted and used as the model input by employing the network embedding techniques. Integrating both educational relation information and topology information in the concept map, CMKT adopts the recurrent neural network to perform knowledge tracing tasks. Comprehensive evaluations conducted on five public educational datasets of four different subjects (more than 8000 learners and their 300 000 records) demonstrate the promise and effectiveness of CMKT: The average area under ROC curve (AUC) and overall prediction accuracy (ACC) achieve 0.82 and 0.75, respectively, and CMKT outperforms all the baselines by at least 12.2% and 9.2% in terms of AUC and ACC.
- 7 January 2021
- Published by Springer Nature in Data Mining and Knowledge Discovery
- Vol. 35 (2) , 533-567
Social explorative attention based recommendation for content distribution platforms
- 7 January 2021
- Published by Springer Nature in Data Mining and Knowledge Discovery
- Vol. 35 (2) , 533-567
Social explorative attention based recommendation for content distribution platforms
- Open Access
Privacy Threats Against Federated Matrix Factorization
Matrix Factorization has been very successful in practical recommendation applications and e-commerce. Due to data shortage and stringent regulations, it can be hard to collect sufficient data to build performant recommender systems for a single company. Federated learning provides the possibility to bridge the data silos and build machine learning models without compromising privacy and security. Participants sharing common users or items collaboratively build a model over data from all the participants. There have been some works exploring the application of federated learning to recommender systems and the privacy issues in collaborative filtering systems. However, the privacy threats in federated matrix factorization are not studied. In this paper, we categorize federated matrix factorization into three types based on the partition of feature space and analyze privacy threats against each type of federated matrix factorization model. We also discuss privacy-preserving approaches. As far as we are aware, this is the first study of privacy threats of the matrix factorization method in the federated learning framework.
Open AccessPrivacy Threats Against Federated Matrix Factorization
Matrix Factorization has been very successful in practical recommendation applications and e-commerce. Due to data shortage and stringent regulations, it can be hard to collect sufficient data to build performant recommender systems for a single company. Federated learning provides the possibility to bridge the data silos and build machine learning models without compromising privacy and security. Participants sharing common users or items collaboratively build a model over data from all the participants. There have been some works exploring the application of federated learning to recommender systems and the privacy issues in collaborative filtering systems. However, the privacy threats in federated matrix factorization are not studied. In this paper, we categorize federated matrix factorization into three types based on the partition of feature space and analyze privacy threats against each type of federated matrix factorization model. We also discuss privacy-preserving approaches. As far as we are aware, this is the first study of privacy threats of the matrix factorization method in the federated learning framework.
- Open Access
Vertex-reinforced Random Walk for Network Embedding
In this paper, we study the fundamental problem of random walk for network embedding. We propose to use non-Markovian random walk, variants of vertex-reinforced random walk (VRRW), to fully use the history of a random walk path. To solve the getting stuck problem of VRRW, we introduce an exploitation-exploration mechanism to help the random walk jump out of the stuck set. The new random walk algorithms share the same convergence property of VRRW and thus can be used to learn stable network embeddings. Experimental results on two link prediction benchmark datasets and three node classification benchmark datasets show that our proposed approach reinforce2vec can outperform state-of-the-art random walk based embedding methods by a large margin.
Open AccessVertex-reinforced Random Walk for Network Embedding
In this paper, we study the fundamental problem of random walk for network embedding. We propose to use non-Markovian random walk, variants of vertex-reinforced random walk (VRRW), to fully use the history of a random walk path. To solve the getting stuck problem of VRRW, we introduce an exploitation-exploration mechanism to help the random walk jump out of the stuck set. The new random walk algorithms share the same convergence property of VRRW and thus can be used to learn stable network embeddings. Experimental results on two link prediction benchmark datasets and three node classification benchmark datasets show that our proposed approach reinforce2vec can outperform state-of-the-art random walk based embedding methods by a large margin.
- 11 November 2019
- Published by Springer Nature in Cognitive Computation
- Vol. 12 (3) , 499-512
Modeling Marked Temporal Point Process Using Multi-relation Structure RNN
- 11 November 2019
- Published by Springer Nature in Cognitive Computation
- Vol. 12 (3) , 499-512
Modeling Marked Temporal Point Process Using Multi-relation Structure RNN
- Open Access
Explainable Fashion Recommendation: A Semantic Attribute Region Guided Approach
In fashion recommender systems, each product usually consists of multiple semantic attributes (e.g., sleeves, collar, etc). When making cloth decisions, people usually show preferences for different semantic attributes (e.g., the clothes with v-neck collar). Nevertheless, most previous fashion recommendation models comprehend the clothing images with a global content representation and lack detailed understanding of users' semantic preferences, which usually leads to inferior recommendation performance. To bridge this gap, we propose a novel Semantic Attribute Explainable Recommender System (SAERS). Specifically, we first introduce a fine-grained interpretable semantic space. We then develop a Semantic Extraction Network (SEN) and Fine-grained Preferences Attention (FPA) module to project users and items into this space, respectively. With SAERS, we are capable of not only providing cloth recommendations for users, but also explaining the reason why we recommend the cloth through intuitive visual attribute semantic highlights in a personalized manner. Extensive experiments conducted on real-world datasets clearly demonstrate the effectiveness of our approach compared with the state-of-the-art methods.
Open AccessExplainable Fashion Recommendation: A Semantic Attribute Region Guided Approach
In fashion recommender systems, each product usually consists of multiple semantic attributes (e.g., sleeves, collar, etc). When making cloth decisions, people usually show preferences for different semantic attributes (e.g., the clothes with v-neck collar). Nevertheless, most previous fashion recommendation models comprehend the clothing images with a global content representation and lack detailed understanding of users' semantic preferences, which usually leads to inferior recommendation performance. To bridge this gap, we propose a novel Semantic Attribute Explainable Recommender System (SAERS). Specifically, we first introduce a fine-grained interpretable semantic space. We then develop a Semantic Extraction Network (SEN) and Fine-grained Preferences Attention (FPA) module to project users and items into this space, respectively. With SAERS, we are capable of not only providing cloth recommendations for users, but also explaining the reason why we recommend the cloth through intuitive visual attribute semantic highlights in a personalized manner. Extensive experiments conducted on real-world datasets clearly demonstrate the effectiveness of our approach compared with the state-of-the-art methods.