TensorFlow Enabled Genetic Programming

TL;DR

This paper demonstrates that integrating TensorFlow with genetic programming significantly accelerates computation, with substantial performance improvements on various datasets and hardware configurations, including CPUs and GPUs.

Contribution

It introduces a TensorFlow-enabled genetic programming approach that achieves large speedups over traditional scalar implementations across diverse datasets and hardware.

Findings

2-15x speedup on small datasets with TensorFlow

875x performance increase on medium datasets using a single CPU core

GPU configurations outperform CPU on large datasets by 1.3x

Abstract

Genetic Programming, a kind of evolutionary computation and machine learning algorithm, is shown to benefit significantly from the application of vectorized data and the TensorFlow numerical computation library on both CPU and GPU architectures. The open source, Python Karoo GP is employed for a series of 190 tests across 6 platforms, with real-world datasets ranging from 18 to 5.5M data points. This body of tests demonstrates that datasets measured in tens and hundreds of data points see 2-15x improvement when moving from the scalar/SymPy configuration to the vector/TensorFlow configuration, with a single core performing on par or better than multiple CPU cores and GPUs. A dataset composed of 90,000 data points demonstrates a single vector/TensorFlow CPU core performing 875x better than 40 scalar/Sympy CPU cores. And a dataset containing 5.5M data points sees GPU configurations out-performing CPU configurations on average by 1.3x.