Multilingual lip sync from audio.
Self-supervised 3D optical flow.
Occlusion aware unsupervised learning of optical flow.
Dynamic computational time for visual attention.
Semantic image segmentation with deep convolutional neural networks.
Object detection with deep convolutional neural networks.
Image captioning and question answering with multimodal recurrent neural networks.
Learning from massive noisy labeled data for visual classification.
Feedback neural networks for object detection and fine-grained recognition