A Survey on Syntax-Guided Program Synthesis

PhD Qualifying Examination


Title: "A Survey on Syntax-Guided Program Synthesis"

by

Mr. Jiarong WU


Abstract:

In recent years, a surge of interests in program synthesis arose from both 
the communities of programming language and machine learning. Behind this 
phenomenon is not only the fast growth of satisfiability modulo theories 
(SMT) solvers and neural network (NN), but also the introduction of 
syntactic constraints. The classical formulation of program synthesis 
problem is to find a program that meets a given logical specification. In 
contrast, syntax-guided synthesis (SyGuS) leverages syntactical templates 
to restrict the search space beyond traditional logical specification, 
thus makes the search space more tractable. An annual SyGuS competition 
has been established to encourage the improvement of SyGuS solvers.

This survey begins with the formulation of syntax-guided synthesis 
problem, followed by several example instances. The majority of this 
survey is the illustrations of different algorithms to tackle the SyGuS 
problem. They are selected from the SyGuS solver competition in the past 
years. In the last part, we discuss the evaluation of different solvers 
and explore opportunities to further improve the solver performance.


Date:			Thursday, 28 May 2020

Time:                  	10:00am - 12:00noon

Zoom meeting:           https://hkust.zoom.us/j/96994112085

Committee Members:	Prof. Shing-Chi Cheung (Supervisor)
  			Dr. Charles Zhang (Chairperson)
 			Dr. Shuai Wang
 			Dr. Raymond Wong


**** ALL are Welcome ****