DagSemProc.08051.5.pdf
- Filesize: 22 kB
- 3 pages
Document
N-gram GP: Early results and half-baked ideas
Authors
-
Part of:
Volume:
Dagstuhl Seminar Proceedings, Volume 8051
Series: Dagstuhl Seminar Proceedings (DagSemProc) - License:
Creative Commons Attribution 4.0 International license
- Publication Date: 2008-05-06
Cite AsGet BibTex
Nicholas Freitag McPhee and Riccardo Poli. N-gram GP: Early results and half-baked ideas. In Theory of Evolutionary Algorithms. Dagstuhl Seminar Proceedings, Volume 8051, pp. 1-3, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2008)
https://doi.org/10.4230/DagSemProc.08051.5
https://doi.org/10.4230/DagSemProc.08051.5
Abstract
In this talk I present N-gram GP, a system for evolving linear GP programs using an EDA style system to update the probabilities of different 3-grams (triplets) of instructions. I then pick apart some of the evolved programs in an effort to better understand the properties of this approach and identify ways that it might be extended.
Doing so reveals that there are frequently cases where the system needs two triples of the form ABC and ABD to solve the problem, but can only choose between them probabilistically in the EDA phase. I present the entirely untested idea of creating a new pseudo-instruction that is a duplicate of a key instruction. This could potentially allow the system to learn, for example, that AB is always followed by C, while AB' is always followed by D.
Keywords
- Genetic programming
- estimation of distribution algorithms
- linear GP
- machine learning
Metrics
- Access Statistics
-
Total Accesses (updated on a weekly basis)
0PDF Downloads