Minjun Zhao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Minjun Zhao has a total of 30 co-authors. Here are the top co-authors by number of shared publications:
- Shiqin Liu
College of Computer Science and Technology, Zhejiang University, Hangzhou, China
7 shared publications - Jiajun Bu
Zhejiang Key Lab of Accessible Perception & Intelligent Systems, Zhejiang University, Hangzhou, China
5 shared publications - Chaozhuo Li
Microsoft Research Asia, Beijing, China
3 shared publications - Litian Zhang
Beijing University of Posts and Telecommunications, Beijing, China
2 shared publications - Lu Chen
Zhejiang University, Hangzhou, China
2 shared publications - Xiaoyi Zhou
Hainan University
2 shared publications
Latest Publications
- Open Access
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 1234-1245
Process In-Context Learning: Confusion-Aware Demonstration Retrieval and Dynamic Insertion for Mathematical Reasoning
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 1234-1245
Open AccessProcess In-Context Learning: Confusion-Aware Demonstration Retrieval and Dynamic Insertion for Mathematical Reasoning
- 9 May 2026
- Published by Springer Nature
- p. 104-120
RICE: Enhancing Transparency in Recommendation via Interaction-Based Counterfactual Explanations
- 9 May 2026
- Published by Springer Nature
- p. 104-120
RICE: Enhancing Transparency in Recommendation via Interaction-Based Counterfactual Explanations
- Open Access
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 1892-1902
LLMCBR: Large Language Model-based Multi-View and Multi-Grained Learning for Bundle Recommendation
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 1892-1902
Open AccessLLMCBR: Large Language Model-based Multi-View and Multi-Grained Learning for Bundle Recommendation
- Open Access
- 27 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 13558-13560
'What Can I Cook?' LetMeCook: An LLM-Based Interactive System for Personalized Recipe Generation
- 27 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 13558-13560
Open Access'What Can I Cook?' LetMeCook: An LLM-Based Interactive System for Personalized Recipe Generation
- Open Access
- 8 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 2278-2287
ModalSync: Synchronizing User Behavior with Multimodal Features for Multimodal Pre-training Recommendation
- 8 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 2278-2287
Open AccessModalSync: Synchronizing User Behavior with Multimodal Features for Multimodal Pre-training Recommendation
- Open Access
- 8 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 2988-2992
Collaborative Optimization Approach for Workflow Agents in User Behavior Modeling
- 8 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 2988-2992
Open AccessCollaborative Optimization Approach for Workflow Agents in User Behavior Modeling
- Open Access
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 9505-9514
Deeply Fusing Semantics and Interactions for Item Representation Learning via Topology-driven Pre-training
Learning item representation is crucial for a myriad of on-line e-commerce applications. The nucleus of retail item representation learning is how to properly fuse the semantics within a single item, and the interactions across different items generated by user behaviors (e.g., co-click or co-view). Product semantics depict the intrinsic characteristics of the item, while the interactions describe the relationships between items from the perspective of human perception. Existing approaches either solely rely on a single type of information or loosely couple them together, leading to hindered representations. In this work, we propose a novel model named TESPA to reinforce semantic modeling and interaction modeling mutually. Specifically, collaborative filtering signals in the interaction graph are encoded into the language models through fine-grained topological pre-training, and the interaction graph is further enriched based on semantic similarities. After that, a novel multi-channel co-training paradigm is proposed to deeply fuse the semantics and interactions under a unified framework. In a nutshell, TESPA is capable of enjoying the merits of both sides to facilitate item representation learning. Experimental results of on-line and off-line evaluations demonstrate the superiority of our proposal.
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 9505-9514
Open AccessDeeply Fusing Semantics and Interactions for Item Representation Learning via Topology-driven Pre-training
Learning item representation is crucial for a myriad of on-line e-commerce applications. The nucleus of retail item representation learning is how to properly fuse the semantics within a single item, and the interactions across different items generated by user behaviors (e.g., co-click or co-view). Product semantics depict the intrinsic characteristics of the item, while the interactions describe the relationships between items from the perspective of human perception. Existing approaches either solely rely on a single type of information or loosely couple them together, leading to hindered representations. In this work, we propose a novel model named TESPA to reinforce semantic modeling and interaction modeling mutually. Specifically, collaborative filtering signals in the interaction graph are encoded into the language models through fine-grained topological pre-training, and the interaction graph is further enriched based on semantic similarities. After that, a novel multi-channel co-training paradigm is proposed to deeply fuse the semantics and interactions under a unified framework. In a nutshell, TESPA is capable of enjoying the merits of both sides to facilitate item representation learning. Experimental results of on-line and off-line evaluations demonstrate the superiority of our proposal.
- Open Access
SparDL: Distributed Deep Learning Training with Efficient Sparse Communication
Top-k sparsification has recently been widely used to reduce the communication volume in distributed deep learning. However, due to the Sparse Gradient Accumulation (SGA) dilemma, the performance of top-k sparsification still has limitations. Recently, a few methods have been put forward to handle the SGA dilemma. Regrettably, even the state-of-the-art method suffers from several drawbacks, e.g., it relies on an inefficient communication algorithm and requires extra transmission steps. Motivated by the limitations of existing methods, we propose a novel efficient sparse communication framework, called SparDL. Specifically, SparDL uses the Spar-Reduce-Scatter algorithm, which is based on an efficient Reduce-Scatter model, to handle the SGA dilemma without additional communication operations. Besides, to further reduce the latency cost and improve the efficiency of SparDL, we propose the Spar-All-Gather algorithm. Moreover, we propose the global residual collection algorithm to ensure fast convergence of model training. Finally, extensive experiments are conducted to validate the superiority of SparDL.
Open AccessSparDL: Distributed Deep Learning Training with Efficient Sparse Communication
Top-k sparsification has recently been widely used to reduce the communication volume in distributed deep learning. However, due to the Sparse Gradient Accumulation (SGA) dilemma, the performance of top-k sparsification still has limitations. Recently, a few methods have been put forward to handle the SGA dilemma. Regrettably, even the state-of-the-art method suffers from several drawbacks, e.g., it relies on an inefficient communication algorithm and requires extra transmission steps. Motivated by the limitations of existing methods, we propose a novel efficient sparse communication framework, called SparDL. Specifically, SparDL uses the Spar-Reduce-Scatter algorithm, which is based on an efficient Reduce-Scatter model, to handle the SGA dilemma without additional communication operations. Besides, to further reduce the latency cost and improve the efficiency of SparDL, we propose the Spar-All-Gather algorithm. Moreover, we propose the global residual collection algorithm to ensure fast convergence of model training. Finally, extensive experiments are conducted to validate the superiority of SparDL.
- 27 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 35 (12) , 12663-12678
Finding Materialized Models for Model Reuse
Materialized model query aims to find the most appropriate materialized model as the initial model for model reuse. It is the precondition of model reuse, and has recently attracted much attention. Nonetheless, the existing methods suffer from the need to provide source data, limited range of applications, and inefficiency since they do not construct a suitable metric to measure the target-related knowledge of materialized models. To address this, we present ${\sf MMQ}$ , a source-data free, general, efficient, and effective materialized model query framework. It uses a Gaussian mixture-based metric called separation degree to rank materialized models. For each materialized model, ${\sf MMQ}$ first vectorizes the samples in the target dataset into probability vectors by directly applying this model, then utilizes Gaussian distribution to fit for each class of probability vectors, and finally uses separation degree on the Gaussian distributions to measure the target-related knowledge of the materialized model. Moreover, we propose an improved ${\sf MMQ}$ ( ${\sf I-MMQ}$ ), which significantly reduces the query time while retaining the query performance of ${\sf MMQ}$ . Extensive experiments on a range of practical model reuse workloads demonstrate the effectiveness and efficiency of ${\sf MMQ}$ .
- 27 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 35 (12) , 12663-12678
Finding Materialized Models for Model Reuse
Materialized model query aims to find the most appropriate materialized model as the initial model for model reuse. It is the precondition of model reuse, and has recently attracted much attention. Nonetheless, the existing methods suffer from the need to provide source data, limited range of applications, and inefficiency since they do not construct a suitable metric to measure the target-related knowledge of materialized models. To address this, we present ${\sf MMQ}$ , a source-data free, general, efficient, and effective materialized model query framework. It uses a Gaussian mixture-based metric called separation degree to rank materialized models. For each materialized model, ${\sf MMQ}$ first vectorizes the samples in the target dataset into probability vectors by directly applying this model, then utilizes Gaussian distribution to fit for each class of probability vectors, and finally uses separation degree on the Gaussian distributions to measure the target-related knowledge of the materialized model. Moreover, we propose an improved ${\sf MMQ}$ ( ${\sf I-MMQ}$ ), which significantly reduces the query time while retaining the query performance of ${\sf MMQ}$ . Extensive experiments on a range of practical model reuse workloads demonstrate the effectiveness and efficiency of ${\sf MMQ}$ .
- 13 September 2018
- Published by Springer Nature
- p. 208-220
A Secure Blind Watermarking Scheme Based on Dual Frequency Domains and Ergodic Matrix
- 13 September 2018
- Published by Springer Nature
- p. 208-220
A Secure Blind Watermarking Scheme Based on Dual Frequency Domains and Ergodic Matrix