Inductive and Deductive Synthesis for Database Applications

Abstract

Program synthesis is a promising method for building efficient, flexible software by deriving low-level implementations from high-level specifications. In this thesis, I use programming-languages techniques to develop systems for synthesizing high-performance, specialized software and to build better general-purpose program-synthesis algorithms. I describe two new synthesis systems. First, I present a full-featured, synthesis-based pipeline for generating database implementations that are specialized to query workloads. This project shows that synthesis is a promising approach for building systems software, but building efficient synthesizers is still difficult, and in general a new synthesizer must be built for every new language. To address this need, I present a new, general-purpose inductive synthesizer, and show that it offers state-of-the-art performance on several challenging tasks.