Computer vision has become ubiquitous in our society, with a variety of applications in image/video search and understanding, medicine, drones, and self-driving cars. As the core to many of the above applications, visual analysis such as image classification, segmentation, localization and detection would be among the well-known problems in computer vision. Recent developments in neural networks (a.k.a. deep learning) approaches have greatly advanced the performance of these state-of-the-art visual recognition systems. This course is a deep dive into the details of deep learning architectures, with a particular focus on understanding and designing learnable models for solving various vision tasks.