Current advances in computing and information sciences.

Deep learning is an area of machine learning that has enabled enormous progress on long-standing problems in visual analytics and machine perception. This course will start with a graduate-level introduction to deep learning, and review neural networks and related theory in machine learning that is needed to understand how deep learning algorithms work. After gaining the prerequisite background knowledge, the class will review the latest deep learning algorithms for computer vision and machine perception. Students will read and present recent papers on visual analytics topics, including eye tracking, image classification, video understanding, model explanation, etc. The course will make an emphasis on approaches with practical relevance, and prepare students to use state-of-the-art deep learning algorithms for processing and understanding highly structured data such as images, videos, and time-series. Students are expected to have programming experience and to be comfortable with probability, linear algebra and calculus. No prior background in machine learning or pattern recognition is required.

An introduction to the underlying concepts of computer vision and image understanding. The course will consider fundamental topics, including image formation, edge detection, texture analysis, color, segmentation, shape analysis, detection of objects in images and high level image representation. Depending on the interest of the class, more advanced topics will be covered, such as image database retrieval or robotic vision. Programming assignments are an integral part of the course. Note: students who complete CSCI-431 may not take CSCI-631 for credit.