Illinois CS+X Program: New Degrees at the Intersection of CS and Everything

A vast number of university undergraduate students need a solid base of computing+data to address challenging problems in social science, applied science, the humanities, policy, business, and the like.  But they do not aspire to be computer (or even data) scientists.  The Illinois CS+X program is a systematic experiment to take CS “wide” into these diverse disciplines. CS+X is a portfolio of novel B.S. degrees, launched in 2014, architected as (Half-CS + Half-X), delivered as a degree in the Dept. of X.  Several degrees are now on offer, ranging from CS+Anthropology to CS+Astronomy.  The program has surprising traction – many partner +X departments now count a significant fraction of their majors as being “CS+X”.  About a dozen more CS+X degrees are now in various stages of design and approval.   A recent talk about our Illinois CS+X program, delivered at the 2016 US National Academies Workshop on the Growth of CS Undergrad Enrollments, is here.

MOOCs: Coursera course, VLSI CAD: Logic to Layout

Since 2013 I have been teaching a   VLSI oriented MOOC on Coursera. This is a version of my class from Carnegie Mellon, moved into the MOOC era, with all the assignments uploaded and evaluated in the cloud.  This was the first EDA (Electronic Design Automation) class offered as a MOOC;  over 50,000 registered learners to date.  Topics include computational Boolean algebra, logic synthesis, technology mapping, timing, placement and routing.  This is high-level tour of the foundational algorithms that make it possible for people to design chips with a billion elements, taught from an algorithms/data-structures sort of angle.  Aimed at folks who want to build tools, and also at people doing real chip designs, who want to know why the tools behave the way they do.  I’ve written and lectured a bit about the experience of teaching EDA “to the planet”, some of these talks and a papers are included here.

Here is the video intro for the VLSI CAD MOOC:

Probabilistic Graphical Models

I have also taught a graduate course on graphical models, using the Koller/Friedman book, and based largely on the lecture notes from Carlos Guestrin (formerly at CMU) and Andrew McCallum (UMASS).  I tend to be rather less mathematics-first in my preferred attack on these topics, though, which gives a different spin to these lectures.  For example, here’s my version of the (rather long) lecture on random sampling, done as PPT with live annotation.Bug me if you’d like to see any of the other lectures.

Curriculum Design Efforts

I’ve been fortunate to be part of several interesting curriculum design efforts,  over the last 25 years of my life in academia.  Here a couple of those historical papers.