About Me

Hi! I am a second-year M.S. student in Systems Science at National University of Defense Technology, advised by Assoc. Prof. Tingjin Luo. Previously, I was an undergraduate student in Information and Computing Science at Shanxi Univeristy.

I work on improving the reliability and safety of deep learning models. During my Master’s program, I am interested in learning robust multi-view models against incomplete data and limited supervision information. Currently, I also research hallucinations in Vision-Language Models, working with Assoc. Prof. Xu Yang from Southeast University.

Links: [CV]

Recent News

  • (08/2024) Create a GitHub Repo to share some Multi-view Multi-class benchmarks.

Selected Publications ($\dagger$ indicates equal contribution.)

Adversarial Graph Fusion for Incomplete Multi-view Semi-supervised Learning with Tensorial Imputation
Zhangqi Jiang, Tingjin Luo, Xu Yang, Xinyan Liang

Coming soon

TL;DR: We present the first framework to tackle the Sub-Cluster Problem in incomplete GMvSSL via a min-max optimization framework and tensor-based structure recovery.

Label Prompts Graph: Incomplete Multi-view Semi-Supervised Classification with Adaptive Graph Regularization
Tingjin Luo$^\dagger$, Zhangqi Jiang$^\dagger$, Chenping Hou

Coming soon

TL;DR: Can we automatically learn optimal partial similarity graphs from incomplete multi-view data with the help of extrinsic supervised signals from labels?

sym

Devils in Middle Layers of Large Vision-Language Models: Interpreting, Detecting and Mitigating Object Hallucinations via Attention Lens
Zhangqi Jiang, Junkai Chen, Beier Zhu, Tingjin Luo, Yankun Shen, Xu Yang

CVPR, 2025 [code]

TL;DR: Delving into how VLMs process visual information from image tokens and how this affects the generation of object hallucinations.

sym

Deep Incomplete Multi-View Learning Network with Insufficient Label Information
Zhangqi Jiang, Tingjin Luo, Xinyan Liang

AAAI, 2024 [code]

TL;DR: To tackle the dual incompleteness of views and labels problem in multi-view scenarios, we leverage graph neural networks for view recovery and a pseudo-labeling strategy for mining supervision information.

sym

Pancancer survival prediction using a deep learning architecture with multimodal representation and integration
Ziling Fan$^{\dagger}$, Zhangqi Jiang$^{\dagger}$, Hengyu Liang, Chao Han

Bioinformatics Advances, 2023 [code]

TL;DR: A deep learning architecture to adaptively integrate multi-omics data (e.g., miRNA and CNV) for pancancer survival prediction.

Projects

Honors and Awards