Xiaodan Du

I am a Ph.D. student at Toyota Technological Institute at Chicago (TTIC), working on computer vision and machine learning. I am fortunate to be advised by Professor Greg Shakhnarovich. My main research interests are generative models, 2D/3D generation and interpretation of deep generative models.

Before my study at TTIC, I have received an MS degree in Computer Science, advised by Professor Svetlana Lazebnik; and a BS degree in Civil Engineering, both from the amazing University of Illinois Urbana-Champaign (UIUC).

I am also an avid collector of 1800s to 1900s European historical and commemorative Medals. Here are a subset of my finest collections.

• News
• Research
• Internship
• Services
• Collection

We introduce SHuBERT, a self-supervised representation learning approach that adapts the masked prediction for multi-stream visual sign language input, learning to predict multiple targets for corresponding to clustered hand, face, and body pose streams. SHuBERT achieves state-of-the-art performance on multiple sign language benchmarks.

We introduce SignMusketeers for Sign Language Translation at Scale. Using only 3% of compute, 41x less pretraining data, and 160x less pretraining epochs, it achieves competitive performance(-0.4 BLEU) compared to the recent ASL-English Translation SOTA.

We introduce Intrinsic LoRA (I-LoRA), a universal, plug-and-play approach that transforms any generative model into a scene intrinsic predictor, capable of extracting intrinsic scene maps directly from the original generator network.

Generating 3D objects from 2D diffusion models by chaining scores with NeRF gradients.

We propose Fast text2StyleGAN, a natural language interface that adapts pre-trained GANs for text-guided human face synthesis.

This website's source code is borrowed from here.