Technical Reports

Display by Author: A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z
Search by for:
Bottom-Up ILP using Large Refinement Steps
Authors: Arias, Marta; Khardon, Roni
Date:April 2004
Download Formats: [PDF]
The system is a bottom up ILP system for learning multi-clause and multi-predicate function free Horn expressions in the framework of learning from interpretations. The paper introduces a new implementation of the same base algorithm which gives several orders of magnitude speedup as well as extending the capabilities of the system. New tools include several fast engines for subsumption tests, handling real valued features, and pruning. We also discuss using data from the standard ILP setting in our framework, which in some cases allows for further speedup. The efficacy of the system is demonstrated on several ILP datasets.

Faculty: for help posting a technical report please visit the User Guide.