CMSC828R

Seminar course reviewing current literature and opportunities in enhancing audio-to-audio and photo-to-photo transformations using deep learning architectures, with the goal of making the output perceptually pleasing for consumption by humans. Student must participate in all classes and lead the discussion in one class. Those taking for 3 credits must fulfill these requirements plus a project lasting at least half the semester, where a suitable algorithm is evaluated and written up in the form of a paper.