ISR
Premkumar Devanbu
No cost to attend.
Click here for directions and parking information.
We are now deep in the era of Big Code. Millions and Billions of SLOC beckon to us, from the firmament of GitHub, BitBucket, SourceForge, StackOverflow and ASF.
It’s the Golden Age of Data. Dogma & Ideology are out, Empiricism is decidedly IN. Why?
Reason 1) Got a theory, a belief, a question or even a dogma about how to engineer software? No need to argue over it: There's enough code, developers, discussions, bugs, and coding history out there: if you can conceive of a (natural) experiment to shed light on your question or dogma, chances are, you can find relevant data.
Reason 2) Got a coding problem? a question about code? a Bug? No need to fret, or run off and try to prove a hairy Hoare triple; It turns out, people write code that is highly repetitive, and statistically predictable; Chances are, you can find something in the billions of SLOC, and billions of changes, out there, that's going to help you out.
In this talk I'll present work at UC Davis (spanning about 10 years now), along both these lines, with a special focus on the latter, where we have developed some exciting new approaches based on techniques adapted from Statistical Natural Language Processing, and deep learning.
Prem Devanbu received his B.Tech from IIT Madras, in Chennai, India, and his PhD from Rutgers University. After working long years as a developer and a researcher at Bell Labs and its various offshoots, he fled New Jersey traffic to join the CS faculty at UC Davis in late 1997, where he bikes to lecture halls. He even has his own web page.