HPC suite

SIBILA

Automated, interpretable machine learning on HPC platforms.

About

SIBILA is an automated machine-learning (AutoML) platform focused on building interpretable predictive models. It handles the full pipeline — preprocessing, model selection, hyperparameter tuning and explanation — and runs natively on HPC infrastructure.

Applications span ADMET prediction, cardiovascular risk modelling, hospital-readmission prediction, drought monitoring, and other tabular forecasting tasks.

Lead authors: Antonio J. Banegas-Luna

Key features

AutoML on HPC

Automated model search optimised for supercomputing platforms.

Interpretability first

Feature attribution and explanation alongside predictions, not as an afterthought.

Multi-domain

Used for ADMET, clinical risk, environmental and other tabular problems.

Access

HPC suite

Source code on GitHub

Open source · private SaaS deployment on demand

https://github.com/bio-hpc/sibila →

Cite this tool

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

Banegas-Luna A.J., Pérez-Sánchez H. · AI 5(4), 2353–2374, 2024

DOI: 10.3390/ai5040116