October 9, 2020
One common issue with program synthesis is that it doesn’t scale: we can synthesize small programs from small specifications, but make either dimension bigger and our tools get stuck. But in many programming domains, small programs are exactly the ones programmers need the most help with. In this talk I’ll present some recent work on synthesizing small, tricky programs for quantized machine learning and for digital signal processors. I’ll also show some tools we can give programmers to help them build specialized synthesizers like these.
About James Bornholt
James Bornholt is an assistant professor of computer science at the University of Texas at Austin. His research is in programming languages and formal methods, with a focus on automated program verification and synthesis, and on their application to systems and architecture.