Tr-dq: Time-rotation diffusion quantization
Y Shao, D Lin, M Yan, S Chen, F Zeng, M Liao, A Ma, Z Yan, H Wang, et al. "Tr-dq: Time-rotation diffusion quantization." AAAI 2026, 40(11), 8869-8877.
CV
Education
- B.S. in Mechanical Engineering, Harbin Engineering University, 2017 – 2021
- M.S. in Electronic Information, University of Chinese Academy of Sciences, 2022 – Present
Research Interests
Image Generation; Video Generation; Multimodal Generation; Model Compression
Publications
Robust Detection in Complex Construction Sites: HiPA-DETR with Weather-Aware and Cross-Domain Generalization
et al., M Zhang (5th author). "Robust Detection in Complex Construction Sites: HiPA-DETR with Weather-Aware and Cross-Domain Generalization." ACM MM 2025 (under review).
Memory Efficient Point Cloud Segmentation with Spatial Group Attention
et al., M Zhang (6th author). "Memory Efficient Point Cloud Segmentation with Spatial Group Attention." ACM MM 2025 (under review).
Vision Also You Need: Navigating Out-of-Distribution Detection with Multimodal Large Language Model
et al., M Zhang (4th author). "Vision Also You Need: Navigating Out-of-Distribution Detection with Multimodal Large Language Model." ICCV 2025 (under review).
AccidentX: A Large-Scale Multimodal BEV Dataset for Traffic Accident Analysis and Prevention
M Zhang et al. "AccidentX: A Large-Scale Multimodal BEV Dataset for Traffic Accident Analysis and Prevention." IROS 2025 (under review).
Controllable Panoramic Video Generation with 360-Degree Motion Consistency for Multiple Control Tasks using a Unified Framework
M Zhang et al. "Controllable Panoramic Video Generation with 360-Degree Motion Consistency for Multiple Control Tasks using a Unified Framework." ICCV 2025 (under review).
AASD: Accelerate Inference by Aligning Speculative Decoding in Multimodal Large Language Models
et al., M Zhang (3rd author). "AASD: Accelerate Inference by Aligning Speculative Decoding in Multimodal Large Language Models." DAC 2025.
PDFT: parameter-diminish fine-tuning for transformer-based models
M Zhang, W Meng, M Jia, J Gu, Y Shao, C Wang, R Xu, Z Ma, X Zhang. "PDFT: parameter-diminish fine-tuning for transformer-based models." The Visual Computer, 41(9), 6745-6755.
PanoDit: Panoramic videos generation with diffusion transformer
M Zhang, Y Chen, R Xu, C Wang, JM Yang, W Meng, J Guo, H Zhao, et al. "PanoDit: Panoramic videos generation with diffusion transformer." AAAI 2025, 39(10), 10040.
Humandreamer: Generating controllable human-motion videos via decoupled generation
B Wang, X Wang, C Ni, G Zhao, Z Yang, Z Zhu, M Zhang, Y Zhou, X Chen, et al. "Humandreamer: Generating controllable human-motion videos via decoupled generation." CVPR 2025, 12391.
Eventvad: Training-free event-aware video anomaly detection
Y Shao, H He, S Li, S Chen, X Long, F Zeng, Y Fan, M Zhang, Z Yan, et al. "Eventvad: Training-free event-aware video anomaly detection." ACM MM 2025, 2586-2595.
AG-SDM: Aquascape generation based on stable diffusion model with low-rank adaptation
M Zhang, J Yang, Y Xian, W Li, J Gu, W Meng, J Zhang, X Zhang. "AG-SDM: Aquascape generation based on stable diffusion model with low-rank adaptation." Computer Animation and Virtual Worlds, 35(3), e2252.
HTCViT: an effective network for image classification and segmentation based on natural disaster datasets
Z Ma, W Li, M Zhang, W Meng, S Xu, X Zhang. "HTCViT: an effective network for image classification and segmentation based on natural disaster datasets." The Visual Computer, 39(8), 3285-3297.
Feaco: Reaching robust feature-level consensus in noisy pose conditions
J Gu, J Zhang, M Zhang, W Meng, S Xu, J Zhang, X Zhang. "Feaco: Reaching robust feature-level consensus in noisy pose conditions." ACM MM 2023, 3628-3636.
Projects
- 2022.09 – 2023.08 Large-Scale Disaster Identification and Localization Equipment (National Project)
- Collaborator: Research Institute of Highway, Ministry of Transport
- Reconstructed disaster scenes from UAV imagery; accurately identified debris, landslides, and flood-affected areas.
- Achieved over 90% accuracy in detecting damaged buildings and vehicles from large-scale aerial images using YOLO.
- Results published in The Visual Computer and accepted at IROS.
Internship
- 2025.02 – Present Algorithm Intern, Beijing Zhipu Huazhang Technology Co., Ltd.
- Responsible for 2D digital human algorithm optimization and deployment.
- Explored virtual live-streaming technology combining panoramic video generation and digital human techniques; work submitted to NeurIPS 2025.
Honors & Awards
- 2023 – 2024 Merit Student, University of Chinese Academy of Sciences
- 2022 – 2023 Third Prize, AI Forum, University of Chinese Academy of Sciences
- 2022 – 2023 Merit Student, University of Chinese Academy of Sciences
- 2020 – 2021 CET-6: 485
- 2022 – 2023 JLPT N2: 159
Student Activities
- 2023 – 2024 Publicity Officer, Graduate Student Association, Institute of Automation, CAS
- 2024 – 2025 Vice Director of Practice Department, Graduate Student Association, Institute of Automation, CAS