Research Topics
Detailed description of the four main research areas and their interconnections.
My research spans machine learning, evolutionary computation, optimization, and computational neuroscience. There is a high degree of overlap between these areas β evolutionary methods are applied to optimize neural architectures, probabilistic graphical models are used in both machine learning and evolutionary algorithms, and ML methods underpin the neuroscience work. Click on any area for more detailed information and publication links.
Machine Learning
Probabilistic graphical models, neural networks, NLP, classification, adversarial ML, explainability, and data augmentation methods.
Evolutionary Computation
Estimation of distribution algorithms (EDAs), genetic programming, neuroevolution, and evolutionary multi-objective optimization.
Optimization
Fitness landscapes, combinatorial and multi-objective optimization, scheduling problems, and RL-based approaches.
Neuroscience
Brain-computer interfaces, unconscious visual processing, decoded neurofeedback, MEG/EEG analysis.
Machine Learning
Probabilistic Graphical Models (PGMs)
Research on learning and inference in graphical models including Bayesian networks, Markov networks, restricted Boltzmann machines, and copula-based models. Applications to unsupervised learning, anomaly detection, and representation learning.
Neural Networks, Deep Learning & Neural Architecture Search
Neural architecture search (NAS), generative models (GANs, VAEs), physics-informed neural networks (PINNs), multi-task learning, semi-supervised classification, and neuroevolutionary approaches.
Natural Language Processing (NLP)
Multi-label hierarchical classification, LLMs for feature selection and anomaly explanation, participation in the ENIA Chair in AI and Language Technology.
Adversarial Machine Learning & Explainability
Generation of adversarial examples in audio and other domains, adversarial attacks on explainable ML methods, uncertainty-aware explanations, and self-explainable neural networks.
Evolutionary Computation
Estimation of Distribution Algorithms (EDAs)
Foundational research on EDAs β learning probabilistic models of promising solutions to guide search. Applications in combinatorial optimization, continuous optimization, and machine learning. Co-author of a key monograph on the topic (Springer, 2002).
Neuroevolution & Neural Architecture Search
Evolutionary methods for optimizing neural network architectures and weights. Factorized NAS models, heterogeneous multi-network models, and neuron-coverage-driven neuroevolutionary algorithms.
Optimization
Scheduling & Combinatorial Optimization
Flexible job-shop scheduling using deep reinforcement learning, constraint programming, offline RL, and behavioral cloning. Real-time scheduling for industrial applications.
Multi-Objective Optimization & Fitness Landscapes
Bi-objective combinatorial optimization, MOEAs, fitness landscape analysis, and benchmarking of continuous multi-objective RL problems.
Neuroscience
Brain Decoding & Neurofeedback
Domain adaptation for brain decoding across imaging paradigms. Development of DecNefLab, a modular simulation framework for decoded neurofeedback research. Work in collaboration with BCBL.
Unconscious Visual Processing & BCI
Investigating properties of unconscious visual information processing in the human brain using ML-based analysis of MEG/EEG data. Brain-computer interface research.