qishuai diao — accepted AI-conference papers

2026

BindWeave: Subject-Consistent Video Generation via Cross-Modal Integration

Zhaoyang Li, Dongjun Qian, Kai Su, qishuai diao, Xiangyang Xia, Chang Liu +3

ICLR 2026poster

Diffusion Transformer has shown remarkable abilities in generating high-fidelity videos, delivering visually coherent frames and rich details over extended durations. However, existing video generation models still fall short in subject-consistent video generation due to an inherent difficulty in pa…

Cited by 0 Source ↗Scholar ↗

2023

Designing BERT for Convolutional Networks: Sparse and Hierarchical Masked Modeling

Keyu Tian, Yi Jiang, qishuai diao, Chen Lin, Liwei Wang, Zehuan Yuan

ICLR 2023top-25%

We identify and overcome two key obstacles in extending the success of BERT-style pre-training, or masked image modeling, to convolutional networks (convnets): (i) convolution operation cannot handle irregular, randomly masked input images; (ii) the single-scale nature of BERT pre-training is incons…

Cited by 138 Source ↗PDF ↗code ↗Scholar ↗