Submission deadline:
January 15, 2025
For submission:
https://www.cec2025.org/index/page.html?id=1298
Scope: Automated Machine Learning (AutoML) optimizes machine learning tasks by automating feature processing, model selection, and parameter tuning, reducing reliance on human expertise. It can match or surpass manual tuning
results, cutting costs and advancing Artificial Intelligence (AI) research. While early AutoML focused on single objectives, real-world applications often require balancing multiple objectives, such as accuracy, model complexity, and resource usage. Evolutionary
Computation (EC) offers promising solutions for Multi-objective AutoML (MO-AutoML).
The aim of this
special session is to gather researchers studying EC for MO-AutoML (EMO-AutoML) to share their research on the following non-exhaustive list of topics:
- Advanced EC techniques on AutoML,
- EC for Multiobjective automated feature engineering, such as feature selection, and feature extraction,
- EC for Multiobjective automated Neural Architecture Search (NAS),
- EC for Multiobjective automated hyperparameter optimization,
- EC for Multiobjective automated data augmentation,
- EC for Multiobjective automated data cleaning,
- EC for Multiobjective model compression,
- EC for Multiobjective model combination,
- EC for Multiobjective ensemble learning,
- EMO-AutoML frameworks,
- EMO-AutoML for different learning tasks such as classification/regression, unsupervised, semi-supervised, self-supervised, few-shot, and transfer learning,
- Real-world applications with MO-AutoML.