The modern field of program synthesis originated as a way to help reduce the effort required to produce programs. In recent years, though, we have started to see program synthesis techniques move beyond traditional problems in software engineering and end user programming and into domains that have been traditionally associated with machine learning and artificial intelligence. In this talk, I will describe how techniques from programming languages can help address problems in machine learning and artificial intelligence, and how the combination of deep learning and program synthesis can open up a new class of neurosymbolic learning techniques with some significant benefits over traditional learning techniques.

Armando Solar-Lezama is a professor of Electrical Engineering and Computer Science and Associate Director and COO of the Computer Science and Artificial Intelligence Lab (CSAIL) at MIT. He is best known for his work on program synthesis and the development of the Sketch program synthesis system. He is currently the lead PI of the NSF funded Expeditions project “Understanding the World through Code” and is also the founder of playskript.com, an online platform for creating interactive presentations.