New information theory course

Assistant Professor Erik Learned-Miller will be teaching a new graduate course, Applied Information Theory, in the fall. He is also teaching a Computer Graphics course this semester.

“Information Theory is finding more and more applications in modern research,” said Learned-Miller. “It is fundamental to signal processing, but is becoming an essential tool in machine learning, networking, and even in such areas as computer vision and databases.” Unlike many traditional Information Theory courses, which are heavily theoretical and proof-oriented, his course will focus on applying information theory to real data and practical problems. In addition to learning basic results about entropy, mutual information, and coding theory, the class will address the statistical issues in estimating these quantities from data samples, a topic often omitted from Information Theory courses. “I hope that this course will give students the basic tools they need to use information theoretic methods in their research, and give them strong intuitions about these very powerful tools,” added Learned-Miller. “I am working currently with Professors Don Towsley, Deepak Ganesan, Micah Adler, and others to make sure that the course addresses the needs of a broad range of research areas.”

The spring semester course (591B) is focused on computer graphics fundamentals from hardware and line-drawing to advanced rendering techniques. While other graphics courses, such as the animation courses taught by Research Associate Professor Beverly Woolf, focus on the use of high-level graphics packages and tools, Learned-Miller’s class will focus on the fundamental concepts that underlie modern graphics techniques. These include the basics of three-dimensional geometry such as transformations, homogeneous coordinates, and perspective; rasterization issues such as sampling and anti-aliasing; hardware issues like analog to digital conversion, accurate color rendering and gamma correction; and some advanced modern techniques such as texture mapping, vertex shading, and optimization methods. “Students are exposed to many different levels of computer architecture and design in their curriculum,” noted Learned-Miller, “from circuits and processors at the lowest level, to operating systems and compilers at the systems level, to programming languages and the high level abstractions of object oriented design. Computer graphics re-emphasizes each level, and shows students how having a thorough knowledge of the entire system gives the engineer flexibility and power to develop innovative and exciting applications. And graphics is also just plain fun. It has always been a favorite of computer science students.”