-
Fitness Approximation through Machine Learning
Authors:
Itai Tzruia,
Tomer Halperin,
Moshe Sipper,
Achiya Elyasaf
Abstract:
We present a novel approach to performing fitness approximation in genetic algorithms (GAs) using machine-learning (ML) models, through dynamic adaptation to the evolutionary state. Maintaining a dataset of sampled individuals along with their actual fitness scores, we continually update a fitness-approximation ML model throughout an evolutionary run. We compare different methods for: 1) switching…
▽ More
We present a novel approach to performing fitness approximation in genetic algorithms (GAs) using machine-learning (ML) models, through dynamic adaptation to the evolutionary state. Maintaining a dataset of sampled individuals along with their actual fitness scores, we continually update a fitness-approximation ML model throughout an evolutionary run. We compare different methods for: 1) switching between actual and approximate fitness, 2) sampling the population, and 3) weighting the samples. Experimental findings demonstrate significant improvement in evolutionary runtimes, with fitness scores that are either identical or slightly lower than that of the fully run GA -- depending on the ratio of approximate-to-actual-fitness computation. Although we focus on evolutionary agents in Gymnasium (game) simulators -- where fitness computation is costly -- our approach is generic and can be easily applied to many different domains.
△ Less
Submitted 21 May, 2024; v1 submitted 6 September, 2023;
originally announced September 2023.
-
A Melting Pot of Evolution and Learning
Authors:
Moshe Sipper,
Achiya Elyasaf,
Tomer Halperin,
Zvika Haramaty,
Raz Lapid,
Eyal Segal,
Itai Tzruia,
Snir Vitrack Tamam
Abstract:
We survey eight recent works by our group, involving the successful blending of evolutionary algorithms with machine learning and deep learning: 1. Binary and Multinomial Classification through Evolutionary Symbolic Regression, 2. Classy Ensemble: A Novel Ensemble Algorithm for Classification, 3. EC-KitY: Evolutionary Computation Tool Kit in Python, 4. Evolution of Activation Functions for Deep Le…
▽ More
We survey eight recent works by our group, involving the successful blending of evolutionary algorithms with machine learning and deep learning: 1. Binary and Multinomial Classification through Evolutionary Symbolic Regression, 2. Classy Ensemble: A Novel Ensemble Algorithm for Classification, 3. EC-KitY: Evolutionary Computation Tool Kit in Python, 4. Evolution of Activation Functions for Deep Learning-Based Image Classification, 5. Adaptive Combination of a Genetic Algorithm and Novelty Search for Deep Neuroevolution, 6. An Evolutionary, Gradient-Free, Query-Efficient, Black-Box Algorithm for Generating Adversarial Instances in Deep Networks, 7. Foiling Explanations in Deep Neural Networks, 8. Patch of Invisibility: Naturalistic Black-Box Adversarial Attacks on Object Detectors.
△ Less
Submitted 8 June, 2023;
originally announced June 2023.
-
EC-KitY: Evolutionary Computation Tool Kit in Python with Seamless Machine Learning Integration
Authors:
Moshe Sipper,
Tomer Halperin,
Itai Tzruia,
Achiya Elyasaf
Abstract:
EC-KitY is a comprehensive Python library for doing evolutionary computation (EC), licensed under the BSD 3-Clause License, and compatible with scikit-learn. Designed with modern software engineering and machine learning integration in mind, EC-KitY can support all popular EC paradigms, including genetic algorithms, genetic programming, coevolution, evolutionary multi-objective optimization, and m…
▽ More
EC-KitY is a comprehensive Python library for doing evolutionary computation (EC), licensed under the BSD 3-Clause License, and compatible with scikit-learn. Designed with modern software engineering and machine learning integration in mind, EC-KitY can support all popular EC paradigms, including genetic algorithms, genetic programming, coevolution, evolutionary multi-objective optimization, and more. This paper provides an overview of the package, including the ease of setting up an EC experiment, the architecture, the main features, and a comparison with other libraries.
△ Less
Submitted 19 April, 2023; v1 submitted 21 July, 2022;
originally announced July 2022.
