References of "Begon, Jean-Michel"
     in
Bookmark and Share    
Full Text
See detailA random walk in Machine Learning
Begon, Jean-Michel ULiege

Conference given outside the academic context (2017)

Since the dawn of machine learning (ML), it hasn’t stop spreading into our everyday lives in new, creative ways. Why Google, Facebook, Amazon and the like have invested so much in ML recently? What can ... [more ▼]

Since the dawn of machine learning (ML), it hasn’t stop spreading into our everyday lives in new, creative ways. Why Google, Facebook, Amazon and the like have invested so much in ML recently? What can (and can’t) ML actually do for us? In this non-technical (and non-exhaustive) talk we will examine ML applications ranging from standard uses to some of the most exotic ones. [less ▲]

Detailed reference viewed: 160 (21 ULiège)
Full Text
Peer Reviewed
See detailGlobally Induced Forest: A Prepruning Compression Scheme
Begon, Jean-Michel ULiege; Joly, Arnaud; Geurts, Pierre ULiege

in Proceedings of Machine Learning Research (2017), 70

Tree-based ensemble models are heavy memory- wise. An undesired state of affairs consider- ing nowadays datasets, memory-constrained environment and fitting/prediction times. In this paper, we propose the ... [more ▼]

Tree-based ensemble models are heavy memory- wise. An undesired state of affairs consider- ing nowadays datasets, memory-constrained environment and fitting/prediction times. In this paper, we propose the Globally Induced Forest (GIF) to remedy this problem. GIF is a fast prepruning approach to build lightweight ensembles by iteratively deepening the current forest. It mixes local and global optimizations to produce accurate predictions under memory constraints in reasonable time. We show that the proposed method is more than competitive with standard tree-based ensembles under corresponding constraints, and can sometimes even surpass much larger models. [less ▲]

Detailed reference viewed: 61 (2 ULiège)
Full Text
Peer Reviewed
See detailJoint learning and pruning of decision forests
Begon, Jean-Michel ULiege; Joly, Arnaud ULiege; Geurts, Pierre ULiege

Conference (2016, September 12)

Decision forests such as Random Forests and Extremely randomized trees are state-of-the-art supervised learning methods. Unfortunately, they tend to consume much memory space. In this work, we propose an ... [more ▼]

Decision forests such as Random Forests and Extremely randomized trees are state-of-the-art supervised learning methods. Unfortunately, they tend to consume much memory space. In this work, we propose an alternative algorithm to derive decision forests under heavy memory constraints. We show that under such constraints our method usually outperforms simpler baselines and can even sometimes beat the original forest. [less ▲]

Detailed reference viewed: 71 (11 ULiège)
Full Text
Peer Reviewed
See detailSLDC: an open-source workflow for object detection in multi-gigapixel images
Mormont, Romain ULiege; Begon, Jean-Michel ULiege; Hoyoux, Renaud ULiege et al

Conference (2016, September 12)

Detailed reference viewed: 143 (47 ULiège)
Full Text
Peer Reviewed
See detailCollaborative analysis of multi-gigapixel imaging data using Cytomine
Marée, Raphaël ULiege; Rollus, Loïc; Stévens, Benjamin et al

in Bioinformatics (2016)

Motivation: Collaborative analysis of massive imaging datasets is essential to enable scientific discoveries. Results: We developed Cytomine to foster active and distributed collaboration of ... [more ▼]

Motivation: Collaborative analysis of massive imaging datasets is essential to enable scientific discoveries. Results: We developed Cytomine to foster active and distributed collaboration of multidisciplinary teams for large-scale image-based studies. It uses web development methodologies and machine learning in order to readily organize, explore, share, and analyze (semantically and quantitatively) multi-gigapixel imaging data over the internet. We illustrate how it has been used in several biomedical applications. Availability: Cytomine (http://www.cytomine.be/) is freely available under an open-source license from http://github.com/cytomine/. A documentation wiki (http://doc.cytomine.be) and a demo server (http://demo.cytomine.be) are also available. [less ▲]

Detailed reference viewed: 183 (32 ULiège)
Full Text
See detailGeneric image classification: random and convolutional approaches
Begon, Jean-Michel ULiege

Master's dissertation (2014)

Supervised learning introduces genericity in the field of image classification, thus enabling fast progress in the domain. Genericity does not imply ease-of-use, however, and the best methods in term of ... [more ▼]

Supervised learning introduces genericity in the field of image classification, thus enabling fast progress in the domain. Genericity does not imply ease-of-use, however, and the best methods in term of accuracy, namely convolutional neural networks, suffer from its lack. In this master thesis, we propose an alternative approach relying on extremely randomized trees and random subwindow extraction combine with elements of the convolutional networks. We explore two modes of utilization of the forest: primarily a direct approach where the forest is the final classifier (ET-DIC) and to a lesser extent, a preprocessing step where the forest is used to build a visual dictionary but where the actual classification is undertaken by a support vector machine (ET-FL). We show that, in both modes, our scheme performs better than without using the convolutional network elements but we are not quite yet reaching their performances. The ET-DIC variant keeps more in the line of classification forest advantages but performs less well as far as accuracy is concerned. This is further highlighted by the remarkable stability of the ET-DIC mode. This stability accounts for the ease-of-use of the method but also prevents elaborated optimization. We were able to score an accuracy of 0.613 whereas the record for this mode without the convolutional network elements was of 0.5367. The ET-FL produces better results at the cost of a greater variability of accuracy due to the loss of the ability to favor the interesting filters and a greater overfitting, consequence of the loss of the ensemble smoothing effect. The accuracies range from 0.55 to 0.7431 depending on the choice of hyper-parameters. The computational cost of both methods is much greater than with a traditional forest, however. [less ▲]

Detailed reference viewed: 32 (5 ULiège)
Full Text
See detailPrototypage d'un serveur de données géographiques maillées: Rasdaman
Begon, Jean-Michel ULiege

Master's dissertation (2011)

In this thesis, we assessed the capability of the Rasdaman software as a full-scale raster data server. In the domain of Geographic Information Systems (GIS), data servers hold an even more prominent ... [more ▼]

In this thesis, we assessed the capability of the Rasdaman software as a full-scale raster data server. In the domain of Geographic Information Systems (GIS), data servers hold an even more prominent position than in traditional information systems. Currently, many solutions, both private and open source, exist for vectorial data and have been extensively tested on practical project. However, solutions for raster data are scarcer. Of the few available solutions, we explored Rasdaman, a rather generic software for multi-dimensional arrays (MDA). As it turns out, Rasdaman has no built-in support for Geographic Coordinate Systems (GCS) and, as such, cannot be used alone, in spite of its MDA language which makes it a strong candidate for web services. Nonetheless, GCS can be build on top of Rasdaman and actually is in the form of a Petascope plugin. All in all, the combination Rasdaman, Petascope and PostGIS make for a powerful and flexible data server for both vector and raster data. [less ▲]

Detailed reference viewed: 23 (4 ULiège)