The 1st International Conference on Evolutionary Computation and Learning (EvoLearn 2027) will take place as part of EvoStar (Evo*).
The fields of Evolutionary Computation (EC) and Machine Learning (ML) have always been tightly intertwined. At the metaphoric level, evolution can be viewed as learning and adapting at the species time scale, while ML acts at the individual scale. At the algorithmic level, both EC and ML fields somehow search for models to fit some available data. Supervised and unsupervised learning aims at fitting available data and generalizing well to unseen data, while reinforcement learning aims at maximizing some reward in a given context. Evolutionary Computation focuses on high-quality regions of a search space defined by a fitness function.
In this context, the incredible blossoming of ML in recent years, which has impacted all research domains, has particularly resonated in the EC community, opening new research opportunities in both fields. The use of ML techniques within EC algorithms, and meta-heuristics at large, is not new, but the tremendous recent progresses of ML have immediatly led to progresses in EC, from online hyperparameter adaptation to powerful surrogate modelling. On the other hand, beyond classical optimization techniques used in ML, there has always been some space for EC methods to successfully tackle problems out of reach of standard techniques, from minimizing non-differentiable losses to optimizing hyperparameters of ML pipelines and to searching rich unstructured search spaces, e.g., in Neural Architecture Search.
Like all other EvoStar events, EvoLearn aims at creating an inclusive, respectful, and open-minded conference environment that invites participants also from other parts of the world independently of political tensions.
Conference Chairs
- Franz Rothlauf
Johannes Gutenberg-Universität Mainz, Germany
rothlauf (at) uni-mainz.de - Marc Schoenauer
Inria, France
marc.schoenauer (at) inria.fr
Publication Chair
- Giorgia Nadizar
University of Toulouse Capitole, France
giorgia.nadizar (at) ut-capitole.fr
Areas of Interest and Contributions
EvoLearn is particularly interested in, but not limited to, original theoretical and/or experimental works combining one way or another Evolutionary Computation and Machine Learning. Some examples are:
- EC for fine-tuning of ML models,
- Evolutionary Reinforcement Learning,
- Evolutionary prompt and pre-prompt optimization,
- Evolutionary Neural Architecture search
- EC or ML for metaheuristic algorithm selection and configuration,
- ML for evolutionary variation operators,
- Representation learning,
- Surrogate modeling,
- Linkage learning.
- Any intricate joint work of EC and ML
Beyond such recombinations, EvoLearn also welcomes contributions to theory, methodology or application of Evolutionary Computation from the point of view of learning at the population level.
Submission Details
Accepted papers will be presented orally or as posters at the event and included in the evo* proceedings, published by Springer Nature in a dedicated volume of the Lecture Notes in Computer Science series. Submissions will be rigorously reviewed for scientific and artistic merit. The reviewing process will be double-blind, so please omit information about the authors in the submitted paper and anonymise links for double-blind review.
Follow these instructions to submit a paper.