FlexGP: Flexible ML with Genetic Programming

Multi-level parallelism provided by FlexGP

The FlexGP framework

FlexGP exploits the multi-level parallelism of GP to tackle large regression problems. FlexGP is composed of a sophisticated learner, a set of launch scripts, and a Java library that provides a TCP/IP communication layer:

  1. Multiple Regression Genetic Programming (MRGP) learner
  2. Decentralized launch protocol
  3. The Java library for P2P communication
  4. Model fusion via Adaptive Regression by Mixing (ARM)

Tutorial

Running FlexGP on Amazon EC2

Publications

Veeramachaneni, K; Arnaldo, I; Derby, O; O’Reilly, UM: FlexGP: Cloud-Based Ensemble Learning with Genetic Programming for Large Regression Problems. Journal of Grid Computing. November, 2014.

Contact

FlexGP is a project of the Any-Scale Learning For All (ALFA) group at MIT. Please contact us at: flexgp@csail.mit.edu

ALFA