Browsing
     by title


0-9 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

or enter first few letters:   
OK
Full Text
See detailMachine Learning Approaches to Power System Security Assessment
Wehenkel, Louis ULg

Post doctoral thesis (1994)

Detailed reference viewed: 32 (6 ULg)
Full Text
Peer Reviewed
See detailMachine Learning based Prediction of Internet Path Dynamics
Wassermann, Sarah ULg; Casas, Pedro; Donnet, Benoît ULg

in ACM CoNEXT Student Workshop: Irvine 12 décembre 2016 (2016, December)

We study the problem of predicting Internet path changes and path performance using traceroute and machine-learning techniques. Path changes are frequently linked to path inflation and performance ... [more ▼]

We study the problem of predicting Internet path changes and path performance using traceroute and machine-learning techniques. Path changes are frequently linked to path inflation and performance degradation. Therefore, predicting their occurrence could improve the analysis of path dynamics using traceroute. By relying on neural networks and using empirical distribution based input features, we show that we are able to predict (i) the remaining life time of a path before it actually changes, and (ii) the number of path changes in a certain time slot with relatively high accuracy. We also show that it is possible to predict path performance in terms of latency, opening the door to novel, machine-learning-based approaches for RTT prediction. [less ▲]

Detailed reference viewed: 42 (2 ULg)
Full Text
See detailMachine Learning Solution Methods for Multistage Stochastic Programming
Defourny, Boris ULg

Doctoral thesis (2010)

This thesis investigates the following question: Can supervised learning techniques be successfully used for finding better solutions to multistage stochastic programs? A similar question had already been ... [more ▼]

This thesis investigates the following question: Can supervised learning techniques be successfully used for finding better solutions to multistage stochastic programs? A similar question had already been posed in the context of reinforcement learning, and had led to algorithmic and conceptual advances in the field of approximate value function methods over the years. This thesis identifies several ways to exploit the combination "multistage stochastic programming/supervised learning" for sequential decision making under uncertainty. Multistage stochastic programming is essentially the extension of stochastic programming to several recourse stages. After an introduction to multistage stochastic programming and a summary of existing approximation approaches based on scenario trees, this thesis mainly focusses on the use of supervised learning for building decision policies from scenario-tree approximations. Two ways of exploiting learned policies in the context of the practical issues posed by the multistage stochastic programming framework are explored: the fast evaluation of performance guarantees for a given approximation, and the selection of good scenario trees. The computational efficiency of the approach allows novel investigations relative to the construction of scenario trees, from which novel insights, solution approaches and algorithms are derived. For instance, we generate and select scenario trees with random branching structures for problems over large planning horizons. Our experiments on the empirical performances of learned policies, compared to golden-standard policies, suggest that the combination of stochastic programming and machine learning techniques could also constitute a method per se for sequential decision making under uncertainty, inasmuch as learned policies are simple to use, and come with performance guarantees that can actually be quite good. Finally, limitations of approaches that build an explicit model to represent an optimal solution mapping are studied in a simple parametric programming setting, and various insights regarding this issue are obtained. [less ▲]

Detailed reference viewed: 176 (18 ULg)
Full Text
See detailMachine learning techniques for atmospheric pollutant monitoring
Sainlez, Matthieu ULg; Heyen, Georges ULg

Poster (2012, January 27)

Machine learning techniques are compared to predict nitrogen oxide (NOx) pollutant emission from the recovery boiler of a Kraft pulp mill. Starting from a large database of raw process data related to a ... [more ▼]

Machine learning techniques are compared to predict nitrogen oxide (NOx) pollutant emission from the recovery boiler of a Kraft pulp mill. Starting from a large database of raw process data related to a Kraft recovery boiler, we consider a regression problem in which we are trying to predict the value of a continuous variable. Generalization is done on the worst case configuration possible to make sure the model is adequate: the training period concerns stationary operations while test periods mainly focus on NOx emissions during transient operations. [less ▲]

Detailed reference viewed: 32 (8 ULg)
Full Text
Peer Reviewed
See detailMachine learning techniques to assess the performance of a gait analysis system
Pierard, Sébastien ULg; Phan-Ba, Rémy; Van Droogenbroeck, Marc ULg

in European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN) (2014, April 24)

This paper presents a methodology based on machine learning techniques to assess the performance of a system measuring the trajectories of the lower limbs extremities for the follow-up of patients with ... [more ▼]

This paper presents a methodology based on machine learning techniques to assess the performance of a system measuring the trajectories of the lower limbs extremities for the follow-up of patients with multiple sclerosis. We show how we have established, with the help of machine learning, four important properties about this system: (1) an automated analysis of gait characteristics provides an improved analysis with respect to that of a human expert, (2) after learning, the gait characteristics provided by this system are valuable compared to measures taken by stopwatches, as used in the standardized tests, (3) the motion of the lower limbs extremities contains a lot of useful information about the gait, even if it is only a small part of the body motion, (4) a measurement system combined with a machine learning tool is sensitive to intra-subject modifications of the walking pattern. [less ▲]

Detailed reference viewed: 92 (15 ULg)
Full Text
See detailMachine Learning to Balance the Load in Parallel Branch-and-Bound
Marcos Alvarez, Alejandro ULg; Wehenkel, Louis ULg; Louveaux, Quentin ULg

E-print/Working paper (2015)

We describe in this paper a new approach to parallelize branch-and-bound on a certain number of processors. We propose to split the optimization of the original problem into the optimization of several ... [more ▼]

We describe in this paper a new approach to parallelize branch-and-bound on a certain number of processors. We propose to split the optimization of the original problem into the optimization of several subproblems that can be optimized separately with the goal that the amount of work that each processor carries out is balanced between the processors, while achieving interesting speedups. The main innovation of our approach consists in the use of machine learning to create a function able to estimate the difficulty (number of nodes) of a subproblem of the original problem. We also present a set of features that we developed in order to characterize the encountered subproblems. These features are used as input of the function learned with machine learning in order to estimate the difficulty of a subproblem. The estimates of the numbers of nodes are then used to decide how to partition the original optimization tree into a given number of subproblems, and to decide how to distribute them among the available processors. The experiments that we carry out show that our approach succeeds in balancing the amount of work between the processors, and that interesting speedups can be achieved with little effort. [less ▲]

Detailed reference viewed: 104 (9 ULg)
See detailMachine learning, neural networks and statistical pattern recognition for voltage security: a comparative study
Wehenkel, Louis ULg; Van Cutsem, Thierry ULg; Pavella, Mania ULg et al

in Proc. 5th International Conference on Intelligent System Applications to Power systems (1994, September)

Detailed reference viewed: 28 (1 ULg)
Peer Reviewed
See detailMachine learning, neural networks and statistical pattern recognition for voltage security: a comparative study
Wehenkel, Louis ULg; Van Cutsem, Thierry ULg; Pavella, Mania ULg et al

in International Journal of Intelligent Systems (1994), 2

Detailed reference viewed: 29 (2 ULg)
Full Text
Peer Reviewed
See detailA Machine Learning-Based Approximation of Strong Branching
Marcos Alvarez, Alejandro ULg; Louveaux, Quentin ULg; Wehenkel, Louis ULg

in INFORMS Journal on Computing (in press)

We present in this paper a new generic approach to variable branching in branch-and-bound for mixed- integer linear problems. Our approach consists in imitating the decisions taken by a good branching ... [more ▼]

We present in this paper a new generic approach to variable branching in branch-and-bound for mixed- integer linear problems. Our approach consists in imitating the decisions taken by a good branching strategy, namely strong branching, with a fast approximation. This approximated function is created by a machine learning technique from a set of observed branching decisions taken by strong branching. The philosophy of the approach is similar to reliability branching. However, our approach can catch more complex aspects of observed previous branchings in order to take a branching decision. The experiments performed on randomly generated and MIPLIB problems show promising results. [less ▲]

Detailed reference viewed: 92 (4 ULg)
Full Text
See detailMachine learning-based feature ranking: Statistical interpretation and gene network inference
Huynh-Thu, Vân Anh ULg

Doctoral thesis (2012)

Machine learning techniques, and in particular supervised learning methods, are nowadays widely used in bioinformatics. Two prominent applications that we target specifically in this thesis are biomarker ... [more ▼]

Machine learning techniques, and in particular supervised learning methods, are nowadays widely used in bioinformatics. Two prominent applications that we target specifically in this thesis are biomarker discovery and regulatory network inference. These two problems are commonly addressed through the use of feature ranking methods that order the input features of a supervised learning problem from the most to the less relevant for predicting the output. This thesis presents, on the one hand, methodological contributions around machine learning-based feature ranking techniques and on the other hand, more applicative contributions on gene regulatory network inference. Our methodological contributions focus on the problem of selecting truly relevant features from machine learning-based feature rankings. Unlike the p-values returned by univariate tests, relevance scores derived from machine learning techniques to rank the features are usually not statistically interpretable. This lack of interpretability makes the identification of the truly relevant features among the top-ranked ones a very difficult task and hence prevents the wide adoption of these methods by practitioners. Our first contribution in this field concerns a procedure, based on permutation tests, that estimates for each subset of top-ranked features the probability for that subset to contain at least one irrelevant feature (called CER for "conditional error rate"). As a second contribution, we performed a large-scale evaluation of several, existing or novel, procedures, including our CER method, that all replace the original relevance scores with measures that can be interpreted in a statistical way. These procedures, which were assessed on several artificial and real datasets, differ greatly in terms of computing times and the tradeoff they achieve in terms of false positives and false negatives. Our experiments also clearly highlight that using model performance as a criterion for feature selection is often counter-productive. The problem of gene regulatory network inference can be formulated as several feature selection problems, each one aiming at discovering the regulators of one target gene. Within this family of methods, we developed the GENIE3 algorithm that exploits feature rankings derived from tree-based ensemble methods to infer gene networks from steady-state gene expression data. In a second step, we derived two extensions of GENIE3 that aim to infer regulatory networks from other types of data. The first extension exploits expression data provided by time course experiments, while the second extension is related to genetical genomics datasets, which contain expression data together with information about genetic markers. GENIE3 was best performer in the DREAM4 In Silico Multifactorial challenge in 2009 and in the DREAM5 Network Inference challenge in 2010, and its extensions perform very well compared to other methods on several artificial datasets. [less ▲]

Detailed reference viewed: 864 (70 ULg)
Peer Reviewed
See detailMachine monstrueuse. Retour sur l'excès, l'exhibition et le gros plan
Jousten, Lison ULg

Conference (2016, October 14)

La communication propose de revenir sur certains aspects monstrueux du cinématographe. À la fois vectrice et créatrice d’un imaginaire intimement associé à la machine, il s’agit principalement d’envisager ... [more ▼]

La communication propose de revenir sur certains aspects monstrueux du cinématographe. À la fois vectrice et créatrice d’un imaginaire intimement associé à la machine, il s’agit principalement d’envisager comment cette « machine à voir » (Viva Paci, 2012) est aussi étroitement liée à la question de la monstruosité. [less ▲]

Detailed reference viewed: 28 (2 ULg)
Full Text
See detailLa machine pénale à travers l’œuvre de José Giovanni
Quintart, Aurélie ULg

Master's dissertation (2015)

Nous avons participé au séminaire Droit et Culture, sous l’égide de M. Nicolas Thirion. Dans ce cadre, nous avons remis un mémoire écrit, mais également présenté celui-ci lors d’un colloque organisé à ... [more ▼]

Nous avons participé au séminaire Droit et Culture, sous l’égide de M. Nicolas Thirion. Dans ce cadre, nous avons remis un mémoire écrit, mais également présenté celui-ci lors d’un colloque organisé à l’ULg le 7 mai 2015. Notre thème était le suivant : La machine pénale à travers l’oeuvre de Jose Giovanni. Notre mémoire était consacré à l’œuvre du réalisateur et écrivain Giovanni, un ancien détenu qui a beaucoup questionné par son œuvre le milieu judiciaire, pénal et carcéral. Nous avons utilisé cette œuvre pour revisiter et expliciter les théories de Lucien François, telles qu’exposées dans « Le Cap des tempêtes ». Notre travail a donc porté sur la philosophie du droit et a permis une réflexion sur l’essence de ce dernier. Nous avons également cherché à mettre en lumière le parallèle, déjà pressenti par M. François, entre criminels et policiers. En effet, notre mémoire visait aussi à démontrer qu’hommes de lois et truands utilisent des techniques de pression et tentent d’influencer l’attitude d’autrui de manière fort similaire. À ce titre, ces différents protagonistes gravitent dans des mondes semblables, caractérisés par la présence d’importants mécanismes et jeux de pouvoir. [less ▲]

Detailed reference viewed: 8 (0 ULg)
Full Text
Peer Reviewed
See detailMachine perfusion in clinical trials : "machine vs. solution effects"
Treckmann, Jürgen; Moers, Cyril; Smits, Jacqueline M et al

in Transplant International (2012), 25

Detailed reference viewed: 22 (0 ULg)
Full Text
Peer Reviewed
See detailMachine Perfusion or cold storage in deceased-donor kidney transplantation
Moers, C.; Smits, J.; Maathuis, M. H. et al

in New England Journal of Medicine [=NEJM] (2009), 360

BACKGROUND Static cold storage is generally used to preserve kidney allografts from deceased donors. Hypothermic machine perfusion may improve outcomes after transplantation, but few sufficiently powered ... [more ▼]

BACKGROUND Static cold storage is generally used to preserve kidney allografts from deceased donors. Hypothermic machine perfusion may improve outcomes after transplantation, but few sufficiently powered prospective studies have addressed this possibility. METHODS In this international randomized, controlled trial, we randomly assigned one kidney from 336 consecutive deceased donors to machine perfusion and the other to cold storage. All 672 recipients were followed for 1 year. The primary end point was delayed graft function (requiring dialysis in the first week after transplantation). Secondary end points were the duration of delayed graft function, delayed graft function defined by the rate of the decrease in the serum creatinine level, primary nonfunction, the serum creatinine level and clearance, acute rejection, toxicity of the calcineurin inhibitor, the length of hospital stay, and allograft and patient survival. RESULTS Machine perfusion significantly reduced the risk of delayed graft function. Delayed graft function developed in 70 patients in the machine-perfusion group versus 89 in the cold-storage group (adjusted odds ratio, 0.57; P = 0.01). Machine perfusion also significantly improved the rate of the decrease in the serum creatinine level and reduced the duration of delayed graft function. Machine perfusion was associated with lower serum creatinine levels during the first 2 weeks after transplantation and a reduced risk of graft failure (hazard ratio, 0.52; P = 0.03). One-year allograft survival was superior in the machine-perfusion group (94% vs. 90%, P = 0.04). No significant differences were observed for the other secondary end points. No serious adverse events were directly attributable to machine perfusion. CONCLUSIONS Hypothermic machine perfusion was associated with a reduced risk of delayed graft function and improved graft survival in the first year after transplantation. (Current Controlled Trials number, ISRCTN83876362.) [less ▲]

Detailed reference viewed: 38 (7 ULg)
Full Text
Peer Reviewed
See detailMachine perfusion versus cold storage for preservation of kidneys from expanded criteria donors after brain death
Treckmann, Jürgen; Moers, Cyril; Smits, Jacqueline M et al

in Transplant International (2011), 24

The purpose of this study was to analyze the possible effects of machine perfusion (MP) versus cold storage (CS) on delayed graft function (DGF) and early graft survival in expanded criteria donor kidneys ... [more ▼]

The purpose of this study was to analyze the possible effects of machine perfusion (MP) versus cold storage (CS) on delayed graft function (DGF) and early graft survival in expanded criteria donor kidneys (ECD). As part of the previously reported international randomized controlled trial 91 consecutive heartbeating deceased ECDs – defined according to the United Network of Organ Sharing definition – were included in the study. From each donor one kidney was randomized to MP and the contralateral kidney to CS. All recipients were followed for 1 year. The primary endpoint was DGF. Secondary endpoints included primary nonfunction and graft survival. DGF occurred in 27 patients in the CS group (29.7%) and in 20 patients in the MP group (22%). Using the logistic regression model MP significantly reduced the risk of DGF compared with CS (OR 0.460, P = 0.047). The incidence of nonfunction in the CS group (12%) was four times higher than in the MP group (3%) (P = 0.04). One-year graft survival was significantly higher in machine perfused kidneys compared with cold stored kidneys (92.3% vs. 80.2%, P = 0.02). In the present study, MP preservation clearly reduced the risk of DGF and improved 1-year graft survival and function in ECD kidneys. [less ▲]

Detailed reference viewed: 23 (2 ULg)
Full Text
Peer Reviewed
See detailMachine perfusion versus cold storage for the preservation of kidneys from donors ≥ years allocated in the Eurotransplant Senior Programme
GALLINAT, Anja; MOERS, Cyril; TRECKMANN, Jürgen et al

in Nephrology Dialysis Transplantation (2012), 27

Background. In the Eurotransplant Senior Programme (ESP), kidneys from donors aged ≥65 years are preferentially allocated locally and transplanted into patients aged ≥65 years on dialysis. The purpose of ... [more ▼]

Background. In the Eurotransplant Senior Programme (ESP), kidneys from donors aged ≥65 years are preferentially allocated locally and transplanted into patients aged ≥65 years on dialysis. The purpose of this study was to analyse whether the results of transplantation in the ESP can be improved by preservation of organs by hypothermic machine perfusion (MP) compared with simple cold storage (CS). Methods. Overall, 85 deceased heart-beating donors ≥65 years of age were included in this analysis with follow-up until 1 year post-transplant. For each donor, one kidney was randomly assigned to preservation by CS and the contralateral kidney to MP from organ procurement until transplantation. Delayed graft function (DGF), primary non-function (PNF) and 1-year patient and graft survival rates were evaluated as primary and secondary endpoints. Results. The median recipient age was 66 years in both groups and the median cold ischaemia time was 11 h for MP and 10.5 h for CS (P = 0.69). The DGF rate was 29.4% for MP and 34.1% for CS (P = 0.58). Only extended duration of cold ischaemia time was an independent risk factor for the development of DGF (odds ratio 1.2, P < 0.0001). PNF was significantly reduced (3.5% MP versus 12.9% CS, P = 0.02). The 1-year patient and graft survival rates were similar for MP and CS (94% versus 95% and 89 versus 81%, P > 0.05). The 1-year graft survival rate was significantly improved after MP in recipients who developed DGF (84% MP versus 48% CS, P = 0.01). Conclusions. Continuous pulsatile hypothermic MP for kidneys from donors aged ≥65 years can reduce the rate of never-functioning kidneys and improve the 1-year graft survival rate of kidneys with DGF. In this small cohort, the known advantage of MP for the reduction of DGF could not be confirmed, possibly due to relatively short cold ischaemia times. [less ▲]

Detailed reference viewed: 28 (2 ULg)
Peer Reviewed
See detailA Machine-Intelligent System for Automatic Target Recognition
Dudgeon, Dan E.; Verly, Jacques ULg; Delanoy, Richard L.

Conference (1990, November)

Detailed reference viewed: 6 (0 ULg)
Full Text
Peer Reviewed
See detailMachine-learning approaches to power-system security assessment
Wehenkel, Louis ULg

in IEEE Expert (1997), 12(5), 60-72

The paper discusses a framework that uses machine learning and other automatic-learning methods to assess power-system security. The framework exploits simulation models in parallel to screen diverse ... [more ▼]

The paper discusses a framework that uses machine learning and other automatic-learning methods to assess power-system security. The framework exploits simulation models in parallel to screen diverse simulation scenarios of a system, yielding a large database. Using data mining techniques, the framework extracts synthetic information about the simulated system's main features from this database [less ▲]

Detailed reference viewed: 15 (2 ULg)
Full Text
Peer Reviewed
See detailA MACHINE-LEARNING CLASSIFIER FOR EPISODIC MIGRAINE BASED ON VISUAL EVOKED GAMMA BAND ACTIVITY
D'Ostilio, Kevin ULg; Lisicki Martinez, Marco ULg; SCHOENEN, Jean ULg et al

in Cephalalgia : An International Journal of Headache (2016), 36(1S), 56

Introduction: Objective and reliable biomarkers of migraine may be of interest for diagnosis and research purposes. Neuroimaging-based machine-learning classifiers are promising but hampered by ... [more ▼]

Introduction: Objective and reliable biomarkers of migraine may be of interest for diagnosis and research purposes. Neuroimaging-based machine-learning classifiers are promising but hampered by availability and cost issues. Conversely, evoked potential are of easy access and affordable. They have provided increasing evidence that sensory information processing is impaired in migraine. We have used gamma band oscillations (GBOs) of visual evoked potentials (VEPs) to compute a machine-learning neural network classifier in episodic migraine. Materials and methods: We analyzed GBOs from VEPs (6x100 responses). Recordings were performed in two matched samples: a training sample composed of 43 migraine patients (EM) and 20 healthy volunteers (HV) and a validating sample of 18 EM and 10 HV. A logistic regression model of the training sample was performed to evaluate the relevance of the predictor variables. Ten neural networks were automatically generated based on late component frequency, n3-p4 and p4-n4 slopes, 1st block n1-p2 amplitude and age. Results: The logistic regression model of the training sample reached a significant classification rate of 79% (EM: 88%; HV: 60%, p¼0.002). The best neural network was able to classify the groups with an accuracy of 73% in the training phase and 89% in the subsequent validation (success rate HV: 90%; EM: 88%). The mean global accuracy within the training and validating samples were 69% (63–78%) and 84% (82–89%). Conclusions: This machine-learning neural network classifier based on visual GBOs provides an accurate and costefficient tool for objective migraine diagnosis. Further training and validation studies with new cohorts are warranted [less ▲]

Detailed reference viewed: 17 (1 ULg)