O
Oleksandr Polozov
Researcher at Microsoft
Publications - 38
Citations - 1829
Oleksandr Polozov is an academic researcher from Microsoft. The author has contributed to research in topics: Parsing & Program synthesis. The author has an hindex of 16, co-authored 35 publications receiving 1241 citations. Previous affiliations of Oleksandr Polozov include University of Washington.
Papers
More filters
Proceedings ArticleDOI
RAT-SQL: Relation-Aware Schema Encoding and Linking for Text-to-SQL Parsers
TL;DR: This work presents a unified framework, based on the relation-aware self-attention mechanism, to address schema encoding, schema linking, and feature representation within a text-to-SQL encoder and achieves the new state-of-the-art performance on the Spider leaderboard.
Proceedings ArticleDOI
FlashMeta: a framework for inductive program synthesis
Oleksandr Polozov,Sumit Gulwani +1 more
TL;DR: The FlashMeta framework implements a novel program synthesis methodology, allowing synthesizer developers to generate an efficient synthesizer from the mere DSL definition (if properties of the DSL operators have been modeled), and found that 10+ existing industrial-quality mass-market applications based on PBE can be cast as instances of D4.
Book
Program Synthesis
TL;DR: Program synthesis is the task of automatically finding a program in the underlying programming language that satises the user intent expressed in the form of some speci�cation as discussed by the authors.
Proceedings ArticleDOI
Learning syntactic program transformations from examples
Reudismam Rolim,Gustavo Soares,Loris D'Antoni,Oleksandr Polozov,Sumit Gulwani,Rohit Gheyi,Ryo Suzuki,Björn Hartmann +7 more
TL;DR: Refazer as mentioned in this paper is a technique for automatically learning program transformations from examples of code edits performed by developers to fix incorrect programming assignment submissions, which can be used as input-output examples to learn program transformations.
Posted Content
Generative Code Modeling with Graphs
TL;DR: The authors use a graph to represent the intermediate state of the generated output of a source code generation procedure, interleaving grammar-driven expansion steps with graph augmentation and neural message passing steps.