Ensemble Learning of Run-Time Prediction Models for Data-Intensive Scientific Workflows

Abstract

Abstract. Workflow applications for in-silico experimentation involve the processing of large amounts of data. One of the core issues for the efficient management of such applications is the prediction of tasks performance. This paper proposes a novel approach that enables the construction models for predicting task's running-times of data-intensive scientific workflows. Ensemble Machine Learning techniques are used to produce robust combined models with high predictive accuracy. Information derived from workflow systems and the characteristics and provenance of the data are exploited to guarantee the accuracy of the models.The proposed approach has been tested on Bioinformatics workflows for Gene Expressions Analysis over homogeneous and heterogeneous computing environments. Obtained results highlight the convenience of using ensemble models in comparison with single/standalone prediction models. Ensemble learning techniques permitted reductions of the prediction error up to 24.9% in comparison with single-model strategies.

Ensemble learning of runtime prediction models for gene-expression analysis workflows

“…The final objective of our approach is the minimization of the human effort when generating the models without trading off the accuracy of predictions. This paper extends the ideas exposed in previous work [15] by providing an new experimental setting and deepening the analysis of results by applying adequate statistical tests.…”

Exploring the Role of Machine Learning in Scientific Workflows: Opportunities and Challenges

“…were used as independent variables to predict the performance as a dependent variable [82]. More advance ML algorithms for runtime prediction have been explored such as Regression Trees [52], neural networks [82], Long short term memory (LSTM) [53], clustering techniques [82], Support Vector Machine (SVM) [83], or combination of several of these methods [84]. Also, ML has been used to predict run-time of workflows over multiple executions [83] (using similar features of executions for Ensemble ML technique), or in an offline learning mode (using parameters such as task resource consumption, system configuration and input data).…”

“…More advance ML algorithms for runtime prediction have been explored such as Regression Trees [52], neural networks [82], Long short term memory (LSTM) [53], clustering techniques [82], Support Vector Machine (SVM) [83], or combination of several of these methods [84]. Also, ML has been used to predict run-time of workflows over multiple executions [83] (using similar features of executions for Ensemble ML technique), or in an offline learning mode (using parameters such as task resource consumption, system configuration and input data). As offline learning methods require the entire dataset to be available for processing in a dynamic environment, where data are continuously generated, new approaches for leveraging ML techniques should be investigated.…”

Learning Running-time Prediction Models for Gene-Expression Analysis Workflows

“…Para lograr tal objetivo este trabajo utiliza la información de desempeño disponible en los registros de ejecución de los WMSs así como la información de proveniencia de los datos en la aplicación. Este trabajo extiende un estudio que hemos llevado a cabo recientemente [13] mediante el análisis detallado de las predicciones realizadas por los modelos combinados, así también como el estudio de la distribución de los errores de predicción individuales. A su vez, unos resultados más completos pueden encontrarse en [14].…”