Multiple Objective Vector-Based Genetic Programming Using Human-Derived Primitives

Abstract

Traditional genetic programming only supports the use of arithmetic and logical operators on scalar features. The GTMOEP (Georgia Tech Multiple Objective Evolutionary Programming) framework builds upon this by also handling feature vectors, allowing the use of signal processing and machine learning functions as primitives, in addition to the more conventional operators. GTMOEP is a novel method for automated, data-driven algorithm creation, capable of outperforming human derived solutions.As an example, GTMOEP was applied to the problem of predicting how long an emergency responder can remain in a hazmat suit before the effects of heat stress cause the user to become unsafe. An existing third-party physics model was leveraged for predicting core temperature from various situational parameters. However, a sustained high heart rate also means that a user is unsafe. To improve performance, GTMOEP was evaluated to predict an expected pull time, computed from both thresholds during human trials.GTMOEP produced dominant solutions in multiple objective space to the performance of predictions made by the physics model alone, resulting in a safer algorithm for emergency responders to determine operating times in harsh environments. The program generated by GTMOEP will be deployed to a mobile application for their use.

TPOT: A Tree-Based Pipeline Optimization Tool for Automating Machine Learning

“…In auto-sklearn, [5] imposed a short and fixed pipeline structure of a data preprocessor, a feature preprocessor, and a model. In another GPbased AutoML system, [22] allowed the GP algorithm to design arbitrarily-shaped pipelines and found that complex pipelines with several preprocessors and models were useful for signal processing problems. Thus, it may be vital to allow AutoML systems to design arbitrarily-shaped pipelines if they are to achieve human-level competitiveness.…”

A System for Accessible Artificial Intelligence

“…Beyond using GP to perform the machine learning itself, recent work has shown that GP can also be harnessed to optimize a sequence of existing data analysis and machine learning operations on a dataset to maximize the predictive performance of the final machine learning model [30,35]. For example, TPOT 4 is an early prototype that uses GP to optimize a sequence of scikit-learn operations for both classification and regression problems [25][26][27], and has been shown to work quite well across a broad range of application domains ranging from epidemiological studies to image classification to time series prediction [23].…”

Otimização e automação da predição precoce do desempenho de alunos que utilizam juízes online: uma abordagem com algoritmo genético

“…3. Construção automática de pipelines de aprendizagem de máquina com algoritmos genéticos [Zutty et al 2015] demonstraram que a otimização com algoritmos genéticos pode superar os seres humanos na busca de um pipeline de aprendizagem de máquina 2 que melhor se encaixa em uma tarefa de aprendizagem supervisionada. Nesse sentido, [Olson et al 2016] propuseram um método para construção automática de pipelines chamado de Treebased Pipeline Optimization Tool (TPOT), no qual a intençãoé criar, inicialmente, uma população inteira de pipelines aleatórios e evoluí-los com operadores de mutações e cruzamentos ao longo de gerações.…”