Abstraction-based interaction model for synthesis

Hila Peleg*, Shachar Itzhaky, Sharon Shoham

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


Program synthesis is the problem of computing from a specification a program that implements it. New and popular variations on the synthesis problem accept specifications in formats that are easier for the human synthesis user to provide: input-output example pairs, type information, and partial logical specifications. These are all partial specification formats, encoding only a fraction of the expected behavior of the program, leaving many matching programs. This transition into partial specification also changes the mode of work for the user, who now provides additional specifications as long as they are unhappy with the result. This turns synthesis into an iterative, interactive process. We present a formal model for interactive synthesis, leveraging an abstract domain of predicates on programs in order to describe the iterative refinement of the specifications and reduction of the candidate program space. We use this model to describe the behavior of several real-world synthesizers. Additionally, we present two conditions for termination of a synthesis session, one hinging only on the properties of the available partial specifications, and the other also on the behavior of the user. Finally, we show conditions for realizability of the user’s intent, and show the limitations of backtracking when it is apparent a session will fail.

Original languageEnglish
Title of host publicationVerification, Model Checking, and Abstract Interpretation - 19th International Conference, VMCAI 2018, Proceedings
EditorsIsil Dillig, Jens Palsberg
PublisherSpringer Verlag
Number of pages24
ISBN (Print)9783319737201
StatePublished - 2018
Event19th International Conference on Verification, Model Checking, and Abstract Interpretation, 2018 - [state] CA, United States
Duration: 7 Jan 20189 Jan 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10747 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference19th International Conference on Verification, Model Checking, and Abstract Interpretation, 2018
Country/TerritoryUnited States
City[state] CA


Dive into the research topics of 'Abstraction-based interaction model for synthesis'. Together they form a unique fingerprint.

Cite this