Machine Learning and Artificial Intelligence

Figure : Sports sciences through data sciences (Louvain-la-Neuve Running Heatmap)

ICTEAM research activities in this field are conducted by ten primary investigators and about fourty researchers. There are two main domains of activity : Machine Learning and Constraint Programming.

Principal Investigators :

Pierre-Antoine Absil, Jean-Charles Delvenne, Yves Deville, Pierre Dupont, John Lee, Siegfried Nijssen, Marco Saerens, Pierre Schaus, Michel Verleysen, Vincent Wertz

Research Labs :

Machine Learning Group, Constraint Group

Research areas :

The research carried out by the UCLouvain Machine Learning Group (MLG) covers both fundamental and applied aspects of machine learning.

Machine learning aims at mining large collection of data and at building models to predict future data. This multidisciplinary field has links to statistics, signal processing, information theory and optimization. It also covers a wide range of applications such as biomedical data analysis, image and video analysis, time series prediction, graph mining, natural language processing, ...

The group specifically addresses the following topics:

  • High-dimensional, functional and non-linear data analysis
  • Feature and model selection
  • Data visualization and manifold learning
  • Bayesian learning
  • Biomedical signal processing and analysis, including ECG, EEG, and respiratory signal analysis, and medical image filtering
  • High-throughput biological data analysis, including microarray data analysis and next-generation sequencing
  • Temporal series prediction, including electrical workload prediction, financial time-series forecasting, networking measurement prediction
  • Automata and Grammar induction with application to software system modeling
  • Structured data analysis, graph mining and collaborative filtering

The UCL Machine Learning Group is organizing, on a yearly basis since 1993, the European Symposium on Artificial Neural Networks - Advances in Computational Intelligence and Learning

Constraint Programming (CP) is a powerful paradigm for modelling and solving complex combinatorial (optimization) problems. It integrates techniques from artificial intelligence, computer science, operational research and optimization. CP separates the modelling of the problem from the search for solutions. It offers high level modelling languages based on constraints. CP proposes two complementary search mechanisms. Standard CP is based on systematic tree search coupled with pruning techniques to remove infeasable solutions. Constraint-Based Local Search (CBLS) allows heuristic search based on the exploration of neighborhoods. The Constraint Group is mainly interested in consistency techniques, integration of CP and CBLS, graph matching, routing problems, applications in networking, ...

Most recent publications

Below are listed the 10 most recent journal articles and conference papers produced in this research area. You also can access all publications by following this link : see all publications.

Journal Articles

1. Yuan, Xinru; Huang, Wen; Absil, Pierre-Antoine; Gallivan, Kyle A. Computing the matrix geometric mean: Riemannian versus Euclidean conditioning, implementation techniques, and a Riemannian BFGS method. In: Numerical Linear Algebra with Applications, Vol. 27, no.5, p. 2321 (2020). doi:10.1002/nla.2321.

2. Coelho, Frederico; Costa, Marcelo; Verleysen, Michel; Braga, Antônio P. LASSO multi-objective learning algorithm for feature selection. In: Soft Computing, Vol. 24, no.4, p. 9 (2020). doi:10.1007/s00500-020-04734-w.

3. Branders, Vincent; Schaus, Pierre; Dupont, Pierre. Identifying gene-specific subgroups: an alternative to biclustering. In: BMC Bioinformatics, Vol. 20, no.625, p. 13 (2019). doi:10.1186/s12859-019-3289-0.

4. Gousenbourger, Pierre-Yves; Massart, Estelle; Absil, Pierre-Antoine. Data Fitting on Manifolds with Composite Bézier-Like Curves and Blended Cubic Splines. In: Journal of Mathematical Imaging and Vision, Vol. 61, no. 5, p. 645-671 (2018). doi:10.1007/s10851-018-0865-2.

5. Kesiman, Made Windu Antara; Valy, Dona; Burie, Jean-christophe; Paulus, Erick; Sunarya, I. Made Gede; Hadi, Setiawan; Sok, Kim Heng; Ogier, Jean-Marc. Southeast Asian palm leaf manuscript images - a review of handwritten text line segmentation methods and new challenges. In: Journal of Electronic Imaging, Vol. 26, no.1, p. 1-15 (2017). doi:10.1117/1.JEI.26.1.011011.

6. Cambier, Léopold; Absil, Pierre-Antoine. Robust low-rank matrix completion by Riemannian optimization. In: SIAM Journal on Scientific Computing, Vol. 38, no. 5, p. S440-S460 (2016). doi:10.1137/15M1025153.

7. Keim, Daniel A.; Munzner, Tamara; Rossi, Fabrice; Verleysen, Michel. Bridging Information Visualization with Machine Learning. In: Dagstuhl Reports, Vol. 5, no. 3, p. 1-27 (2015). doi:10.4230/DagRep.5.3.1.

8. Frénay, Benoît; Verleysen, Michel. Reinforced Extreme Learning Machines for Fast Robust Regression in the Presence of Outliers. In: IEEE Transactions on Cybernetics, no. 99, p. 13 (17/12/2015). doi:10.1109/TCYB.2015.2504404.

9. De Visscher, Ruben; Delouille, Véronique; Dupont, Pierre; Deledalle, Charles-Alban. Supervised classification of solar features using prior information. In: Journal of Space Weather and Space Climate, Vol. 5, no.A34, p. 1-12 (2015). doi:10.1051/swsc/2015033.

10. Paul, Jérôme; D'Ambrosio, Roberto; Dupont, Pierre. Kernel methods for heterogeneous feature selection. In: Neurocomputing, Vol. 169, p. 187-195 (2015). doi:10.1016/j.neucom.2014.12.098.

Conference Papers

1. Hamer, Victor; Dupont, Pierre. Joint optimization of predictive performance and selection stability. In: ESANN 2020 - Proceedings. Vol. 1, no.1, p. 381-386 (2020). 2020 xxx.

2. Hamer, Victor; Dupont, Pierre. Explicit Control of Feature Relevance and Selection Stability Through Pareto Optimality. In: CEUR Workshop Proceedings. Vol. 2444, no.1, p. 64-79 (2019). CEUR, 2019 xxx.

3. Nguyen, Thanh Son; Gousenbourger, Pierre-Yves; Massart, Estelle; Absil, Pierre-Antoine. Online balanced truncation for linear time-varying systems using continuously differentiable interpolation on Grassmann manifold. In: 2019 6th International Conference on Control, Decision and Information Technologies (CoDIT), IEEE, 2019, 9781728105215, p. 165-170 xxx. doi:10.1109/codit.2019.8820675.

4. Massart, Estelle; Hendrickx, Julien; Absil, Pierre-Antoine. Curvature of the Manifold of Fixed-Rank Positive-Semidefinite Matrices Endowed with the Bures–Wasserstein Metric. In: Lecture Notes in Computer Science : Geometric Science of Information (Geometric Science of information), Frank Nielsen, Frédéric Barbaresco Eds. 2019, 9783030269791, p. 739-748 xxx. doi:10.1007/978-3-030-26980-7_77.

5. Renard, Emilie; Absil, Pierre-Antoine; Gallivan, Kyle A.. Minimax center to extract a common subspace from multiple datasets. In: ESANN 2019 Proceedings, 2019, 978-287-587-065-0, p. 275-280 xxx.

6. Dong, Shuyu; Absil, Pierre-Antoine; Gallivan, Kyle A.. Preconditioned conjugate gradient algorithms for graph regularized matrix completion. In: ESANN 2019 Proceedings, 2019, 978-287-587-065-0, p. 239-244 xxx.

7. Renard, Emilie; Gallivan, Kyle A.; Absil, Pierre-Antoine. A Grassmannian Minimum Enclosing Ball Approach for Common Subspace Extraction. In: Latent Variable Analysis and Signal Separation : Lecture Notes in Computer Science, Springer International Publishing, 2018, 9783319937632, p. 69-78 xxx. doi:10.1007/978-3-319-93764-9_7.

8. Hamer, Victor; Dupont, Pierre. Learning Computationally Efficient Metrics for Large Scale Person Identification. In: Proceedings of the Annual Machine Learning Conference of Belgium and the Netherlands 2018. Vol. /, no./, p. / (2018). 2018 xxx.

9. Dong, Shuyu; Absil, Pierre-Antoine; Gallivan, Kyle. Graph learning for regularized low-rank matrix completion. In: MTNS 2018. p. 460-467 (2018). 2018 xxx.

10. Valy, Dona; Verleysen, Michel; Chhun, Sophea; Burie, Jean-Christophe. A New Khmer Palm Leaf Manuscript Dataset for Document Analysis and Recognition - SleukRith Set. 2017 xxx.