Qingyun Sun
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Qingyun Sun has a total of 80 co-authors. Here are the top co-authors by number of shared publications:
- Jianxin Li
Beihang University, Beijing, China
23 shared publications - Xingcheng Fu
Guangxi Normal University, Guilin, China
13 shared publications - Hao Peng
Beihang University, Beijing, China
7 shared publications - Philip S. Yu
University of Illinois Chicago
7 shared publications - Cheng Ji
SKLCCSE, School of Computer Science and Engineering, Beihang University, China
6 shared publications - Haonan Yuan
SKLCCSE, School of Computer Science and Engineering, Beihang University, Beijing, China
6 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 3259-3268
So-Called Input Transformation-Based Attacks Are Actually Model Transformation-Based Attacks
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 3259-3268
Open AccessSo-Called Input Transformation-Based Attacks Are Actually Model Transformation-Based Attacks
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 9539-9548
DualG-MRAG: Decoupling Macro-Reasoning and Micro-Matching for Multimodal Retrieval-Augmented Generation
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 9539-9548
Open AccessDualG-MRAG: Decoupling Macro-Reasoning and Micro-Matching for Multimodal Retrieval-Augmented Generation
- Open Access
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 5653-5662
BotUmc: LLM-Enhanced Uncertainty-Aware Graph Learning for Bot Detection
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 5653-5662
Open AccessBotUmc: LLM-Enhanced Uncertainty-Aware Graph Learning for Bot Detection
- Open Access
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 1431-1441
Rich-Media Re-Ranker: A User Satisfaction-Driven LLM Re-ranking Framework for Rich-Media Search
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 1431-1441
Open AccessRich-Media Re-Ranker: A User Satisfaction-Driven LLM Re-ranking Framework for Rich-Media Search
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 38 (10) , 6302-6320
A Survey on Foundation Models for Structured Data: Tabular, Time Series, and Graphs
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 38 (10) , 6302-6320
A Survey on Foundation Models for Structured Data: Tabular, Time Series, and Graphs
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 13437-13438
2nd Workshop on Frontiers in Graph Machine Learning for the Large Model Era
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 13437-13438
Open Access2nd Workshop on Frontiers in Graph Machine Learning for the Large Model Era
- Open Access
- 11 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 276-282
Timestep-Aware Optimization Dynamics Analysis for Diffusion Training
Multimodal diffusion models have become a central approach for high-quality content generation. Some recent studies explore cross-timestep gradient relations to guide improvements in diffusion training. However, commonly used full-gradient cosine similarity can be distorted by parameter heterogeneity and layer-wise scale differences, leading to misleading estimates of the coupling among timestep-specific objectives. This paper revisits cross-timestep gradient measurement and proposes a corrected protocol for multimodal diffusion training. The corrected measurement shows that gradient similarity decays rapidly with timestep distance, and gradients from distant timesteps become nearly orthogonal as training progresses. These results suggest that diffusion training consists of weakly coupled timestep-wise objectives. They further imply that directly applying history-accumulating mechanisms across timesteps may introduce interference among weakly related timestep objectives. Our study refines the empirical basis for timestep-wise optimization dynamics analysis and provides guidance for optimization design in multimodal content generation.
- 11 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 276-282
Open AccessTimestep-Aware Optimization Dynamics Analysis for Diffusion Training
Multimodal diffusion models have become a central approach for high-quality content generation. Some recent studies explore cross-timestep gradient relations to guide improvements in diffusion training. However, commonly used full-gradient cosine similarity can be distorted by parameter heterogeneity and layer-wise scale differences, leading to misleading estimates of the coupling among timestep-specific objectives. This paper revisits cross-timestep gradient measurement and proposes a corrected protocol for multimodal diffusion training. The corrected measurement shows that gradient similarity decays rapidly with timestep distance, and gradients from distant timesteps become nearly orthogonal as training progresses. These results suggest that diffusion training consists of weakly coupled timestep-wise objectives. They further imply that directly applying history-accumulating mechanisms across timesteps may introduce interference among weakly related timestep objectives. Our study refines the empirical basis for timestep-wise optimization dynamics analysis and provides guidance for optimization design in multimodal content generation.
- Open Access
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1554-1565
Mitigating Privacy Risks in Graph Condensation from a Hyperbolic Geometry Perspective
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1554-1565
Open AccessMitigating Privacy Risks in Graph Condensation from a Hyperbolic Geometry Perspective
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1398-1409
GraphRAG-R1: Graph Retrieval-Augmented Generation with Process-Constrained Reinforcement Learning
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1398-1409
Open AccessGraphRAG-R1: Graph Retrieval-Augmented Generation with Process-Constrained Reinforcement Learning
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 626-637
RAG-GFM: Overcoming In-Memory Bottlenecks in Graph Foundation Models via Retrieval-Augmented Generation
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 626-637
Open AccessRAG-GFM: Overcoming In-Memory Bottlenecks in Graph Foundation Models via Retrieval-Augmented Generation