Blogs (9) >>
SPLASH 2016
Sun 30 October - Fri 4 November 2016 Amsterdam, Netherlands
Wed 2 Nov 2016 14:45 - 15:10 at Matterhorn 2 - Program Synthesis Chair(s): Martin Odersky

Data filtering in spreadsheets is a common problem faced by millions of end-users. The task of data filtering requires a computational model that can separate intended positive and negative string instances. We present a system, FIDEX, that can efficiently learn desired data filtering expressions from a small set of positive and negative string examples.

There are two key ideas of our approach. First, we design an expressive DSL to represent disjunctive filter expressions needed for several real-world data filtering tasks. Second, we develop an efficient synthesis algorithm for incrementally learning consistent filter expressions in the DSL from very few positive and negative examples. A DAG-based data structure is used to succinctly represent a large number of filter expressions, and two corresponding operators are defined for algorithmically handling positive and negative examples, namely, the intersection and subtraction operators. FIDEX is able to learn data filters for 452 out of 460 real-world data filtering tasks in real time (0.22s), using only 2.2 positive string instances and 2.7 negative string instances on average.

Wed 2 Nov

splash-2016-oopsla
13:30 - 15:10: OOPSLA - Program Synthesis at Matterhorn 2
Chair(s): Martin OderskyEPFL, Switzerland
splash-2016-oopsla147808980000013:30 - 13:55
Talk
DOI Pre-print Media Attached
splash-2016-oopsla147809130000013:55 - 14:20
Talk
Venkatesh SrinivasanUniversity of Wisconsin - Madison, Tushar SharmaUniversity of Wisconsin - Madison, USA, Thomas RepsUniversity of Wisconsin - Madison and Grammatech Inc.
DOI Pre-print Media Attached
splash-2016-oopsla147809280000014:20 - 14:45
Talk
Pavel PanchekhaUniversity of Washington, Emina TorlakUniversity of Washington
DOI Media Attached
splash-2016-oopsla147809430000014:45 - 15:10
Talk
Xinyu WangUT Austin, Sumit GulwaniMicrosoft Research, Rishabh SinghMicrosoft Research
DOI Media Attached