SIBILA - Automated Machine-Learning-Based Development of Interpretable Machine-Learning Models on High-Performance Computing Platforms

Aug 4, 2026

Upload by Asher

As machine learning reshapes industries, efficient development and reliable interpretability become critical for deploying models on high-performance computing (HPC) platforms. SIBILA presents an AutoML approach tailored to HPC environments, enabling users to specify objectives and preferences before an automated search selects effective ML pipelines. The framework accelerates pipeline search and evaluation by exploiting HPC parallelism, while enforcing interpretability requirements for regulation and stakeholder understanding. Validated on public datasets, SIBILA attains competitive accuracy with reduced computational overhead. It is provided as a free web service.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages22

Reading Time55 min

AccessFree to download

Introduction

Motivation for efficient interpretable ML development

Proposed Approach: SIBILA

AutoML workflow tailored for HPC

Interpretability-aware model development

Experimental Validation

Public datasets evaluation

Accuracy and computational overhead results

Availability and Use

Free web service for practitioners

Article SIBILA: Automated Machine-Learning-Based Development of Interpretable Machine-Learning Models on High-Performance Computing Platforms Antonio Jesús Banegas-Luna *,† and Horacio Pérez-Sánchez † Citation: Banegas-Luna, A.J.; Pérez-Sánchez, H. SIBILA: Automated Machine-Learning-Based Development of Interpretable Machine-Learning Models on High-Performance Computing Platforms. AI 2024, 5, 2353–2374 . [https://doi.org/10.3390/ai5040116](https://doi.org/10.3390/ai5040116) Academic Editor: Gianni D’Angelo Received: 27 September 2024 Revised: 2 November 2024 Accepted: 11 November 2024 Published: 14 November 2024 Copyright: © 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://](https://)[ ](https://)[creativecommons.org/licenses/by/](creativecommons.org/licenses/by/)[ ](creativecommons.org/licenses/by/)[4.0/](4.0/)) . Structural Bioinformatics and High-Performance Computing (BIO-HPC), Campus de los Jerónimos, Universidad Católica de Murcia (UCAM), Guadalupe, 30107 Murcia, Spain; [hperez@ucam.edu](hperez@ucam.edu) * [Correspondence: ajbanegas@ucam.edu](Correspondence: ajbanegas@ucam.edu); Tel.: +34-968278-821 † These authors contributed equally to this work. Abstract: As machine learning (ML) transforms industries, the need for efficient model development tools using high-performance computing (HPC) and ensuring interpretability is crucial. This paper presents SIBILA, an AutoML approach designed for HPC environments, focusing on the interpretation of ML models. SIBILA simplifies model development by allowing users to set objectives and preferences before automating the search for optimal ML pipelines. Unlike traditional AutoML frameworks, SIBILA is specifically designed to exploit the computational capabilities of HPC platforms, thereby accelerating the model search and evaluation phases. The emphasis on interpretability is particularly crucial when model transparency is mandated by regulations or desired for stakeholder understanding. SIBILA has been validated in different tasks with public datasets. The results demonstrate that SIBILA consistently produces models with competitive accuracy while significantly reducing computational overhead. This makes it an ideal choice for practitioners seeking efficient and transparent ML solutions on HPC infrastructures. SIBILA is a major advancement in AutoML, addressing the rising demand for explainable ML models on HPC platforms. Its integration of interpretability constraints alongside automated model development processes marks a substantial step forward in bridging the gap between computational efficiency and model transparency in ML applications. The tool is available as a web service at no charge. Keywords: explainable machine learning; data fusion; automated machine learning; high-performance computing; deep learning; consensus 1. Introduction The rapid development of technologies has helped artificial intelligence (AI) become a well-known and reliable tool for researchers in academia and industry [1] . Its ability to analyze vast amounts of data has become a powerful tool in science and business [2] . Looking for repetitive patterns among such data collections is a complex but necessary task that needs to be done to extract knowledge from past events. By exploring several samples, AI models can learn the internal relationships among data and use that information to forecast future events or unexplored samples. Machine learning (ML) and its subtype, deep learning (DL), are two typical approaches to AI [3] . Both types of model are flexible enough to analyze a range of datasets, including tabular data, text, time series, and images. This adaptability to different contexts has propelled their application into traditional and fundamental areas of science, such as biology [4,5], chemistry [6,7], and medicine

FAQs

What problem does SIBILA address in machine learning model development?

SIBILA targets the need to develop machine-learning pipelines efficiently on HPC platforms while maintaining interpretability. It automates pipeline search while ensuring interpretability is preserved for compliance and stakeholder understanding.

How does SIBILA differ from traditional AutoML frameworks?

Unlike generic AutoML tools, SIBILA is designed to exploit HPC computational capabilities to speed up the model search and evaluation phases. This reduces the time and compute needed compared with non-HPC-first workflows.

What benefits do the results show for SIBILA?

Experiments on public datasets show that SIBILA produces models with competitive accuracy while significantly reducing computational overhead. This combination supports efficient and transparent ML deployment.

Aug 4, 2026

Upload by Asher