Cancelled Classes (tentative):
September 11 2014 (Distinguished Lecture in Oklahoma)
Note that some of these classes may be covered by Guest Lectures.
Yet, many of the computational problems we face are incomprehensibly intractable: satisfiability of logical formulas, geometry of optimal tours, whole genome (genotypic/haplotypic) assembly, machine learning for probabilistic graphical models, Compressive sensing, etc., to name a few. Despite theoretical pessimism, there have been some spectacular successes (NP-Easy problems).
Even more frustratingly, our ability to understand and formulate new problems that the society faces (with increasing access to computing and ability to generate data) has been dismal. Ideas based on hypotheses generation, prototyping minimal experiments (MVPs), and scaling continue to expose us to enormous risks (and rewards).
This course will focus on computational thinking, problem formulating, characterizing complexity, and problem solving using various heuristic approaches.
A detailed syllabus for the class will be designed in collaboration with the students attending the class.