Publications

2026

Diversifying Similar Subjects for Text-to-image Synthesis with Self-cross Diffusion Guidance and Reward

W. Qiu, J. Wang, Z. Gu, V. T. Hu, M. Tang

IJCV - International Journal on Computer Vision (Accepted)

Contrastive Conditional-Unconditional Alignment For Long-tailed Diffusion Model

F. Chen, A. Villa, G. Liang, L. Fuxin, X. Lu, M. Tang

ECCV 2026 — European Conference on Computer Vision

Paper
Pursuing Minimal Sufficiency in Spatial Reasoning

Y. Guo, Y. Hou, W. Ma, M. Tang, M. H. Yang

ICLR 2026 — International Conference on Learning Representations

Paper Code
Recent Advancements in Deep Learning for SAR-to-Optical Image Translation: A Systematic Literature Review (2021–2025)

T. Dove, Z. Xie, M. Tang, Y. Zhang, X. Ye, N. Jacobs

JSTARS - Journal of Selected Topics in Applied Earth Observations and Remote Sensing

Paper

2025

Self-Cross Diffusion Guidance for Text-to-Image Synthesis of Similar Subjects

W. Qiu, J. Wang, M. Tang

CVPR 2025 — IEEE/CVF Conference on Computer Vision and Pattern Recognition

Project Paper Code
ReDistill: Residual Encoded Distillation for Peak Memory Reduction

F. Chen, G. Datta, M. A. Rafi, H. Jeon, M. Tang

Transactions on Machine Learning Research (TMLR)

Paper Code

2024

Latent Space Editing in Transformer-Based Flow Matching

V. T. Hu, W. Zhang, M. Tang, P. Mettes, D. Zhao, C. Snoek

AAAI 2024 — Proceedings of the AAAI Conference on Artificial Intelligence, 38(3), 2247–2255

Paper Code
ScribbleGen: Generative Data Augmentation Improves Scribble-Supervised Semantic Segmentation

J. Schnell, J. Wang, L. Qi, V. T. Hu, M. Tang

CVPR 2024 Workshop - Workshop on Syntagen at IEEE Conference on Computer Vision and Pattern Recognition

Paper Code
Training Class-Imbalanced Diffusion Model Via Overlap Optimization

D. Yan, L. Qi, V. T. Hu, M. H. Yang, M. Tang

arXiv preprint arXiv:2402.10821

Paper

2023

Decepticon: Attacking Secrets of Transformers

M. Al Rafi, Y. Feng, F. Yao, M. Tang, H. Jeon

IISWC 2023 — IEEE International Symposium on Workload Characterization, pp. 128–139

Paper