Guided Search Space Genetic Programming for identifying energy aware microarchitectural designs

Genetic Programming (GP) is being proposed as a machine learning technique in design space exploration. An evolutionary but heuristic approach by default, GP basically searches the whole input space for suboptimal values, which often translates into long convergence times, more processing and thus inefficient resource utilization. We propose in this paper a Guided Search Space GP (GSS-GP) approach that improves convergence time and accuracy because of the limited search space it uses and the fitness function designed to account for the class disproportionality. Experimental results to identify energy aware microarchitectural designs show the merit of GSS-GP and motivate follow on research.

[1]  Malcolm I. Heywood,et al.  GP Classification under Imbalanced Data sets: Active Sub-sampling and AUC Approximation , 2008, EuroGP.

[2]  David M. Brooks,et al.  Accurate and efficient regression modeling for microarchitectural performance and power prediction , 2006, ASPLOS XII.

[3]  Emiliano Carre.  Evolution of Classification Rules for Comprehensible Knowledge Discovery , 2007 .

[4]  Sotiris B. Kotsiantis,et al.  Combining Bagging, Boosting and Dagging for Classification Problems , 2007, KES.

[5]  María José del Jesús,et al.  A novel genetic cooperative-competitive fuzzy rule based learning method using genetic programming for high dimensional problems , 2008, 2008 3rd International Workshop on Genetic and Evolving Systems.

[6]  Mauro Olivieri.  A genetic approach to the design space exploration of superscalar microprocessor architectures , 2001, ISCAS 2001. The 2001 IEEE International Symposium on Circuits and Systems (Cat. No.01CH37196).

[7]  Hamoud I. Aljamaan,et al.  An empirical study of bagging and boosting ensembles for identifying faulty classes in object-oriented software , 2009, 2009 IEEE Symposium on Computational Intelligence and Data Mining.

[8]  Debashis Ghosh,et al.  Feature selection and molecular classification of cancer using genetic programming. , 2007, Neoplasia.

[9]  Mengjie Zhang,et al.  Fitness Functions in Genetic Programming for Classification with Unbalanced Data , 2007, Australian Conference on Artificial Intelligence.

[10]  Kevin Skadron,et al.  Predictive design space exploration using genetically programmed response surfaces , 2008, 2008 45th ACM/IEEE Design Automation Conference.

[11]  Michael L. Gargano,et al.  Data Mining Using Hybrid Evolutionary Models for Creating Data Classification Rules , 2005 .