Research @ NCRA
Research @ NCRA
UCD NCRA researchers undertake both basic and applied research in a number of application areas, including Financial Modelling, Genetic Programming, Architecture & Design, Music & Sound Synthesis, Computer Graphics & Animation, Social Programming, Combinatorial Optimisation, Adaptive Systems, Bioinformatics and Engineering.
EDGE - Evolution in Dynamic Environments with Grammatical Evolution
The primary objective of the EDGE project is to analyse and improve the ability of Grammatical Evolution (GE) to find solutions in dynamic environments, facilitating its application to hard real-world dynamic problems. Many of the most challenging problems facing researchers and decision makers are those with a dynamic nature. That is, the environment in which the solution exists, and consequently the optimal solution itself, changes over time. The process of evolution has been particularly successful at producing organisms that can survive and adapt to ever-changing environments. Inspired by the workings of biological processes, this project develops a novel method and associated sofware tools which can be applied to solve hard dynamic problems. The project is funded under an SFI Principal Investigator award (08/IN.1/I1868).
EDGE project highlights include:
A collaboration with Bell Labs has resulted in peer-reviewed publications (e.g., GECCO 2011 and GECCO 2012) and a patent application to the European Patent Office (EP2012360002).
Citi Media produced a video featuring our collaboration on Pattern Recognition for Anti Money Laundering. Also available from the IDA Ireland website.
We've published the second book on GE "Foundations in GE for Dynamic Environments" (Further details and book launch photos).
11 journal articles and 42 peer-reviewed conference papers have been published arising from EDGE research (as of August 2013). Sample list provided below.
4 PhD students, and 3 post doctoral researchers have been trained.
O'Neill was interviewed about the research in the EDGE project by Claire O'Connell (Irish Times contributor and freelance journalist).
Our research and collaboration with Bell Labs was highlighted in the Irish Times article "When software takes its cues from nature" (January 2013).
GameBrains is a translational research spin-off of GE in collaboration with the NDRC, Neat Sciences and former Electronic Arts VP Les Edgar.
Two open source software libraries have arisen, namely GEVA (Grammatical Evolution in Java) and GEM (Grammatical Evolution in Matlab).
We hosted the ACM Genetic and Evolutionary Computation Conference GECCO 2011 in Dublin, Ireland 12-16 July 2011, which attracted approximately 600 delegates from 52 countries with an acceptance rate of 38%. A photo diary is available. We are very grateful to Science Foundation Ireland and Failte Ireland for providing financial support, which resulted in GECCO netting the Irish exchequr approximately €1Million.
EDGE project research has won four best paper awards, namely:
We have won prizes in a number of competitions with our Evolved Art:
EDGE funded by: