Short Bio

I am a Senior Researcher at Microsoft GenAI. We are working on OpenAI and Microsoft model training. I received my Ph.D. degree at University of Science and Technology of China (USTC) in 2023. Prior to that, I received my Bachelor degree from the University of Science and Technology of China in 2018.

My research interests focus on multi-model models and general representation learning.

πŸ”₯ News

  • 2025.06: Β πŸŽ‰πŸŽ‰ Being selected as the World’s Top 2% Scientist.
  • 2025.09: Β πŸŽ‰πŸŽ‰ HiFlow was accepted by NeurIPS 2025.
  • 2025.06: Β πŸŽ‰πŸŽ‰ MIR, SAM2Long, Visual-RFT, MM-IFEngine, X-Prompt, Bootstrap3D, and Light-A-Video were accepted by ICCV 2025.
  • 2025.05: Β πŸŽ‰πŸŽ‰ SongComposer was accepted by ACL 2025.
  • 2025.05: Β πŸŽ‰πŸŽ‰ IXC-2.5-Reward and Light-ColPali were accepted by Findings of ACL 2025.
  • 2025.05: Β πŸŽ‰πŸŽ‰ VideoRoPE (Oral) and SongGen were accepted by ICML 2025.
  • 2025.03: Β πŸŽ‰πŸŽ‰ MMStar is selectd as NeurIPS 2024 top10 Influential paper.
  • 2025.02: Β πŸŽ‰πŸŽ‰ Dispider, ViCo, OVO-Bench, and ByTheWay were accepted by CVPR 2025.
  • 2025.01: Β πŸŽ‰πŸŽ‰ MIA-DPO and MotionClone were accepted by ICLR 2025.
  • 2024.11: Β πŸŽ‰πŸŽ‰ ShareGPT4V is selectd as ECCV 2024 top10 Influential paper.
  • 2024.11: Β πŸŽ‰πŸŽ‰ InternLM-XComposer2 and InternVL is selectd as Arxiv2024 CV top10 Influential paper.
  • 2024.09: Β πŸŽ‰πŸŽ‰ ShareGPT4Video, MMDU, and MMLongBench-Doc are accepted by NeurIPS2024 D&B.
  • 2024.09: Β πŸŽ‰πŸŽ‰ InternLM-XComposer2-4KHD, MMStar, and Video-Streaming are accepted by NeurIPS2024.
  • 2024.07: Β πŸŽ‰πŸŽ‰ ShareGPT4V and Long-CLIP are accepted by ECCV2024.
  • 2024.05: Β πŸŽ‰πŸŽ‰ PeCo is selectd as AAAI2023 top10 Influential paper.
  • 2024.02: Β πŸŽ‰πŸŽ‰ OPERA is accepted by CVPR2024 as Highlight.
  • 2023.11: Β πŸŽ‰πŸŽ‰ PointCAT is accepted by TIP.
  • 2023.07: Β πŸŽ‰πŸŽ‰ ELP and RobustMAE are accepted by ICCV2023.
  • 2023.03: Β πŸŽ‰πŸŽ‰ MaskCLIP and DAM-VP are accepted by CVPR2023.
  • 2023.01: Β πŸŽ‰πŸŽ‰ CSWin is selectd as CVPR2022 top10 Influential paper.
  • 2022.12: Β πŸŽ‰πŸŽ‰ PeCo is accepted by AAAI2023.
  • 2022.10: Β πŸŽ‰πŸŽ‰ MaskCLIP won the 1st place in ICinW Industry Track.
  • 2022.07: Β πŸŽ‰πŸŽ‰ BootMAE and CD-Net are accepted by ECCV2022.
  • 2022.03: Β πŸŽ‰πŸŽ‰ CSWin, ICT, Mobile-Former, SI-Adv are accepted by CVPR2022.
  • 2021.11: Β πŸŽ‰πŸŽ‰ PeCo reaches the highest ImageNet-1K accuracy w/o addition data.
  • 2021.08: Β πŸŽ‰πŸŽ‰ Tacr-net is accepted by ACM MM.
  • 2020.09: Β πŸŽ‰πŸŽ‰ GreedyFool is accepted by NeurIPS2020.
  • 2020.03: Β πŸŽ‰πŸŽ‰ Sup-ADV, GVG and LG-GAN are accepted by CVPR2020.
  • 2019.07: Β πŸŽ‰πŸŽ‰ MAN is accepted by ICCV2019.

πŸ“ Selected Publications

CVPR 2023
sym

MaskCLIP: Masked Self-Distillation Advances Contrastive Language-Image Pretraining

Xiaoyi Dong, Jianmin Bao, Yinglin Zheng, Ting Zhang, Dongdong Chen, Hao Yang, Ming Zeng, Weiming Zhang, Lu Yuan, Dong Chen, Fang Wen, Nenghai Yu

CVPR 2023 1st in ICinW Industry Track

AAAI 2023
sym

Peco: Perceptual codebook for bert pre-training of vision transformers

Xiaoyi Dong, Jianmin Bao, Ting Zhang, Dongdong Chen, Weiming Zhang, Lu Yuan, Dong Chen, Fang Wen, Nenghai Yu

AAAI 2023 SOTA ImageNet-1K accuracy

ECCV 2022
sym

Bootstrapped Masked Autoencoders for Vision BERT Pretraining

Xiaoyi Dong, Jianmin Bao, Ting Zhang, Dongdong Chen, Weiming Zhang, Lu Yuan, Dong Chen, Fang Wen, Nenghai Yu

ECCV 2022 | Github

CVPR 2022
sym

CSWin Transformer: A General Vision Transformer Backbone with Cross-Shaped Windows

Xiaoyi Dong, Jianmin Bao, Dongdong Chen, Weiming Zhang, Nenghai Yu, Lu Yuan, Dong Chen, Baining Guo

CVPR 2022 CVPR2022 top10 Influential paper | Github

CVPR 2022
sym

Protecting Celebrities from DeepFake with Identity Consistency Transformer

Xiaoyi Dong, Jianmin Bao, Dongdong Chen, Ting Zhang, Weiming Zhang, Nenghai Yu, Dong Chen, Fang Wen, Baining Guo

CVPR 2022 | Github

NeurIPS 2020
sym

GreedyFool: Distortion-Aware Sparse Adversarial Attack

Xiaoyi Dong, Dongdong Chen, Jianmin Bao, Chuan Qin, Lu Yuan, Weiming Zhang, Nenghai Yu, Dong Chen

NeurIPS 2020 | Github

CVPR 2020
sym

Robust Superpixel-Guided Attentional Adversarial Attack

Xiaoyi Dong, Jiangfan Han, Dongdong Chen, et al., Nenghai Yu

CVPR 2020 | Github

CVPR 2020
sym

Self-Robust 3D Point Recognition via Gather-vector Guidance

Xiaoyi Dong, Dongdong Chen, et al., Nenghai Yu

CVPR 2020

ICCV 2019
sym

Once a MAN: Towards Multi-Target Attack via Learning Multi-Target Adversarial Network Once

Jiangfan Han, Xiaoyi Dong, Ruimao Zhang, Dongdong Chen, Weiming Zhang, Nenghai Yu, Ping Luo, Xiaogang Wang (* Equal Contribution)

ICCV 2019

πŸŽ– Honors and Awards

  • Outstanding Reviewer of CVPR2023, ICCV2021
  • 2023.10, President Scholarship, Chinese Academy of Sciences.
  • 2023.05, Outstanding Doctoral Dissertation Award, USTC
  • 2023.05, Excellent award, Stars of Tomorrow Internship Program, Microsoft Research Asia (MSRA).
  • 2020.06, National Scholarship (The highest scholarship awarded by the Ministry of Education, China).
  • 2019.09, IJCAI-2019 Alibaba Adversarial AI Challenge(AAAC2019). 1st prize in Defense Track, 2nd prize in Non-Target Attack Track.
  • 2018.10, Competition on Adversarial Attacks and Defenses (CAAD2018). 2nd prize in Non-Target Attack Track.

πŸ“– Educations

  • 2018.06 - 2023.06, Ph.D., University of Science and Technology of China.
  • 2014.09 - 2018.06, Undergraduate, University of Science and Technology of China.

Professional Activities

  • Reviewer for CVPR 2021, 2022, 2023
  • Reviewer for ICCV 2021, 2023
  • Reviewer for ECCV 2022
  • Reviewer for NeurIPS 2023
  • Reviewer for AAAI 2020, 2021, 2022
  • Reviewer for IEEE Transactions on Pattern Analysis and Machine Intelligence(TPAMI)
  • Reviewer for IEEE Transactions on Image Processing(TIP)