Modélisation des émotions de l’apprenant et interventions implicites pour les systèmes tutoriels intelligents

La modélisation de l’expérience de l’utilisateur dans les Interactions Homme-Machine est un enjeu important pour la conception et le développement des systèmes adaptatifs intelligents. Dans ce contexte, une attention particulière est portée sur les réactions émotionnelles de l’utilisateur, car elles ont une influence capitale sur ses aptitudes cognitives, comme la perception et la prise de décision. La modélisation des émotions est particulièrement pertinente pour les Systèmes Tutoriels Émotionnellement Intelligents (STEI). Ces systèmes cherchent à identifier les émotions de l’apprenant lors des sessions d’apprentissage, et à optimiser son expérience d’interaction en recourant à diverses stratégies d’interventions. Cette thèse vise à améliorer les méthodes de modélisation des émotions et les stratégies émotionnelles utilisées actuellement par les STEI pour agir sur les émotions de l’apprenant. Plus précisément, notre premier objectif a été de proposer une nouvelle méthode pour détecter l’état émotionnel de l’apprenant, en utilisant différentes sources d’informations qui permettent de mesurer les émotions de façon précise, tout en tenant compte des variables individuelles qui peuvent avoir un impact sur la manifestation des émotions. Pour ce faire, nous avons développé une approche multimodale combinant plusieurs mesures physiologiques (activité cérébrale, réactions galvaniques et rythme cardiaque) avec des variables individuelles, pour détecter une émotion très fréquemment observée lors des sessions d’apprentissage, à savoir l’incertitude. Dans un premier lieu, nous avons identifié les indicateurs physiologiques clés qui sont associés à cet état, ainsi que les caractéristiques individuelles qui contribuent à sa manifestation. Puis, nous avons développé des modèles prédictifs permettant de détecter automatiquement cet état à partir des différentes variables analysées, à travers l’entrainement d’algorithmes d’apprentissage machine. Notre deuxième objectif a été de proposer une approche unifiée pour reconnaître simultanément une combinaison de plusieurs émotions, et évaluer explicitement l’impact de ces émotions sur l’expérience d’interaction de l’apprenant. Pour cela, nous avons développé une plateforme hiérarchique, probabiliste et dynamique permettant de suivre les changements émotionnels de l'apprenant au fil du temps, et d’inférer automatiquement la tendance générale qui caractérise son expérience d’interaction à savoir : l’immersion, le blocage ou le décrochage. L’immersion correspond à une expérience optimale : un état dans lequel l'apprenant est complètement concentré et impliqué dans l’activité d’apprentissage. L’état de blocage correspond à une tendance d’interaction non optimale où l'apprenant a de la difficulté à se concentrer. Finalement, le décrochage correspond à un état extrêmement défavorable où l’apprenant n’est plus du tout impliqué dans l’activité d’apprentissage. La plateforme proposée intègre trois modalités de variables diagnostiques permettant d’évaluer l’expérience de l’apprenant à savoir : des variables physiologiques, des variables comportementales, et des mesures de performance, en combinaison avec des variables prédictives qui représentent le contexte courant de l’interaction et les caractéristiques personnelles de l'apprenant. Une étude a été réalisée pour valider notre approche à travers un protocole expérimental permettant de provoquer délibérément les trois tendances ciblées durant l’interaction des apprenants avec différents environnements d’apprentissage. Enfin, notre troisième objectif a été de proposer de nouvelles stratégies pour influencer positivement l’état émotionnel de l’apprenant, sans interrompre la dynamique de la session d’apprentissage. Nous avons à cette fin introduit le concept de stratégies émotionnelles implicites : une nouvelle approche pour agir subtilement sur les émotions de l’apprenant, dans le but d’améliorer son expérience d’apprentissage. Ces stratégies utilisent la perception subliminale, et plus précisément une technique connue sous le nom d’amorçage affectif. Cette technique permet de solliciter inconsciemment les émotions de l’apprenant, à travers la projection d’amorces comportant certaines connotations affectives. Nous avons mis en œuvre une stratégie émotionnelle implicite utilisant une forme particulière d’amorçage affectif à savoir : le conditionnement évaluatif, qui est destiné à améliorer de façon inconsciente l’estime de soi. Une étude expérimentale a été réalisée afin d’évaluer l’impact de cette stratégie sur les réactions émotionnelles et les performances des apprenants.

Using Neural Network Classifier Support Vector Machine Regression for the prediction of Melting Point of Drug – like compounds

In our study we use a kernel based classification technique, Support Vector Machine Regression for predicting the Melting Point of Drug – like compounds in terms of Topological Descriptors, Topological Charge Indices, Connectivity Indices and 2D Auto Correlations. The Machine Learning model was designed, trained and tested using a dataset of 100 compounds and it was found that an SVMReg model with RBF Kernel could predict the Melting Point with a mean absolute error 15.5854 and Root Mean Squared Error 19.7576

Model selection approaches for non-linear system identification: a review

The identification of non-linear systems using only observed finite datasets has become a mature research area over the last two decades. A class of linear-in-the-parameter models with universal approximation capabilities have been intensively studied and widely used due to the availability of many linear-learning algorithms and their inherent convergence conditions. This article presents a systematic overview of basic research on model selection approaches for linear-in-the-parameter models. One of the fundamental problems in non-linear system identification is to find the minimal model with the best model generalisation performance from observational data only. The important concepts in achieving good model generalisation used in various non-linear system-identification algorithms are first reviewed, including Bayesian parameter regularisation and models selective criteria based on the cross validation and experimental design. A significant advance in machine learning has been the development of the support vector machine as a means for identifying kernel models based on the structural risk minimisation principle. The developments on the convex optimisation-based model construction algorithms including the support vector regression algorithms are outlined. Input selection algorithms and on-line system identification algorithms are also included in this review. Finally, some industrial applications of non-linear models are discussed.

Measuring time series predictability using support vector regression

Most studies involving statistical time series analysis rely on assumptions of linearity, which by its simplicity facilitates parameter interpretation and estimation. However, the linearity assumption may be too restrictive for many practical applications. The implementation of nonlinear models in time series analysis involves the estimation of a large set of parameters, frequently leading to overfitting problems. In this article, a predictability coefficient is estimated using a combination of nonlinear autoregressive models and the use of support vector regression in this model is explored. We illustrate the usefulness and interpretability of results by using electroencephalographic records of an epileptic patient.

Machine Translation and the Disruption of Foreign Language Learning Activities

This study examines the question of how language teachers in a highly technologyfriendly university environment view machine translation and the implications that this has for the personal learning environments of students. It brings an activity-theory perspective to the question, examining the ways that the introduction of new tools can disrupt the relationship between different elements in an activity system. This perspective opens up for an investigation of the ways that new tools have the potential to fundamentally alter traditional learning activities. In questionnaires and group discussions, respondents showed general agreement that although use of machine translation by students could be considered cheating, students are bound to use it anyway, and suggested that teachers focus on the kinds of skills students would need when using machine translation and design assignments and exams to practice and assess these skills. The results of the empirical study are used to reflect upon questions of what the roles of teachers and students are in a context where many of the skills that a person needs to be able to interact in a foreign language increasingly can be outsourced to laptops and smartphones.

Machine translation and the disruption of foreign language learning activities

This study examines the question of how language teachers in a highly technology-friendly university environment view machine translation and the implications that this has for the personal learning environments of students. It brings an activity-theory perspective to the question, examining the ways that the introduction of new tools can disrupt the relationship between different elements in an activity system. This perspective opens up for an investigation of the ways that new tools have the potential to fundamentally alter traditional learning activities. In questionnaires and group discussions, respondents showed general agreement that although use of machine translation by students could be considered cheating, students are bound to use it anyway, and suggested that teachers focus on the kinds of skills students would need when using machine translation and design assignments and exams to practice and assess these skills. The results of the empirical study are used to reflect upon questions of what the roles of teachers and students are in a context where many of the skills that a person needs to be able to interact in a foreign language increasingly can be outsourced to laptops and smartphones.

Low cost non-intrusive home energy monitoring

This thesis presents a low cost non-intrusive home energy monitor built on top of Non-Intrusive Load Monitoring (NILM) concepts and techniques. NILM solutions are already considered low cost alternatives to the big majority of existing commercial energy monitors but the goal here is to make its cost even lower by using a mini netbook as a whole in one solution. The mini netbook is installed in the homes main circuit breaker and computes power consumption by reading current and voltage from the built-in sound card. At the same time, feedback to the users is provided using the 11’’ LCD screen as well as other built-in I/O modules. Our meter is also capable of detecting changes in power and tries to find out which appliance lead to that change and it is being used as part of an eco-feedback platform that was build to study the long terms of energy eco-feedback in individuals. In this thesis the steps that were taken to come up with such a system are presented, from the basics of AC power measurements to the implementation of an event detector and classifier that was used to disaggregate the power load. In the last chapter results from some validation tests that have been performed are presented in order to validate the experiment. It is believed that such a system will not only be important as an energy monitor, but also as an open system than can be easily changed to accommodate and test new or existing nonintrusive load monitoring techniques.

Particle competition and cooperation to prevent error propagation from mislabeled data in semi-supervised learning

Semi-supervised learning is applied to classification problems where only a small portion of the data items is labeled. In these cases, the reliability of the labels is a crucial factor, because mislabeled items may propagate wrong labels to a large portion or even the entire data set. This paper aims to address this problem by presenting a graph-based (network-based) semi-supervised learning method, specifically designed to handle data sets with mislabeled samples. The method uses teams of walking particles, with competitive and cooperative behavior, for label propagation in the network constructed from the input data set. The proposed model is nature-inspired and it incorporates some features to make it robust to a considerable amount of mislabeled data items. Computer simulations show the performance of the method in the presence of different percentage of mislabeled data, in networks of different sizes and average node degree. Importantly, these simulations reveals the existence of the critical points of the mislabeled subset size, below which the network is free of wrong label contamination, but above which the mislabeled samples start to propagate their labels to the rest of the network. Moreover, numerical comparisons have been made among the proposed method and other representative graph-based semi-supervised learning methods using both artificial and real-world data sets. Interestingly, the proposed method has increasing better performance than the others as the percentage of mislabeled samples is getting larger. © 2012 IEEE.

Combined active and semi-supervised learning using particle walking temporal dynamicscombined active and semi-supervised learning using particle walking temporal dynamics

Both Semi-Supervised Leaning and Active Learning are techniques used when unlabeled data is abundant, but the process of labeling them is expensive and/or time consuming. In this paper, those two machine learning techniques are combined into a single nature-inspired method. It features particles walking on a network built from the data set, using a unique random-greedy rule to select neighbors to visit. The particles, which have both competitive and cooperative behavior, are created on the network as the result of label queries. They may be created as the algorithm executes and only nodes affected by the new particles have to be updated. Therefore, it saves execution time compared to traditional active learning frameworks, in which the learning algorithm has to be executed several times. The data items to be queried are select based on information extracted from the nodes and particles temporal dynamics. Two different rules for queries are explored in this paper, one of them is based on querying by uncertainty approaches and the other is based on data and labeled nodes distribution. Each of them may perform better than the other according to some data sets peculiarities. Experimental results on some real-world data sets are provided, and the proposed method outperforms the semi-supervised learning method, from which it is derived, in all of them.

Learning kernels for support vector machines with polynomial powers of sigmoid

In the pattern recognition research field, Support Vector Machines (SVM) have been an effectiveness tool for classification purposes, being successively employed in many applications. The SVM input data is transformed into a high dimensional space using some kernel functions where linear separation is more likely. However, there are some computational drawbacks associated to SVM. One of them is the computational burden required to find out the more adequate parameters for the kernel mapping considering each non-linearly separable input data space, which reflects the performance of SVM. This paper introduces the Polynomial Powers of Sigmoid for SVM kernel mapping, and it shows their advantages over well-known kernel functions using real and synthetic datasets.

Spatial Learning and Stress Response of Male Rats Prenatally Exposed to Dexamethasone

Dexamethasone is routinely administered to women at risk for a preterm birth in order to enhance fetal lung development and reduce uterine contractions. Research has demonstrated possible behavioral abnormalities in adulthood as a result of dexamethasone treatment. Using nonlinear mixed effects modeling, this study found thatprenatal dexamethasone treatment impaired spatial learning and memory of adult male Sprague-Dawley rats. Prenatal dexamethasone treatment also led to more anxiety related behaviors on Elevated Plus Maze testing 1.5 hours after a stress challenge. Because theassumptions underlying the independent samples t-test were violated, the randomization test was used to compare groups on the Elevated Plus Maze.

Linear latent force models using Gaussian Processes

Purely data-driven approaches for machine learning present difficulties when data are scarce relative to the complexity of the model or when the model is forced to extrapolate. On the other hand, purely mechanistic approaches need to identify and specify all the interactions in the problem at hand (which may not be feasible) and still leave the issue of how to parameterize the system. In this paper, we present a hybrid approach using Gaussian processes and differential equations to combine data-driven modeling with a physical model of the system. We show how different, physically inspired, kernel functions can be developed through sensible, simple, mechanistic assumptions about the underlying system. The versatility of our approach is illustrated with three case studies from motion capture, computational biology, and geostatistics.

PV power forecast using a nonparametric PV model

Forecasting the AC power output of a PV plant accurately is important both for plant owners and electric system operators. Two main categories of PV modeling are available: the parametric and the nonparametric. In this paper, a methodology using a nonparametric PV model is proposed, using as inputs several forecasts of meteorological variables from a Numerical Weather Forecast model, and actual AC power measurements of PV plants. The methodology was built upon the R environment and uses Quantile Regression Forests as machine learning tool to forecast AC power with a confidence interval. Real data from five PV plants was used to validate the methodology, and results show that daily production is predicted with an absolute cvMBE lower than 1.3%.

Debonding identification in FRP-plated RC structures using PZT sensors

The application of the Electro-Mechanical Impedance (EMI) method for damage detection in Structural Health Monitoring has noticeable increased in recent years. EMI method utilizes piezoelectric transducers for directly measuring the mechanical properties of the host structure, obtaining the so called impedance measurement, highly influenced by the variations of dynamic parameters of the structure. These measurements usually contain a large number of frequency points, as well as a high number of dimensions, since each frequency range swept can be considered as an independent variable. That makes this kind of data hard to handle, increasing the computational costs and being substantially time-consuming. In that sense, the Principal Component Analysis (PCA)-based data compression has been employed in this work, in order to enhance the analysis capability of the raw data. Furthermore, a Support Vector Machine (SVM), which has been widespread used in machine learning and pattern recognition fields, has been applied in this study in order to model any possible existing pattern in the PCAcompress data, using for that just the first two Principal Components. Different known non-damaged and damaged measurements of an experimental tested beam were used as training input data for the SVM algorithm, using as test input data the same amount of cases measured in beams with unknown structural health conditions. Thus, the purpose of this work is to demonstrate how, with a few impedance measurements of a beam as raw data, its healthy status can be determined based on pattern recognition procedures.

Learning 3D structure from 2D images using LBP features

An automatic machine learning strategy for computing the 3D structure of monocular images from a single image query using Local Binary Patterns is presented. The 3D structure is inferred through a training set composed by a repository of color and depth images, assuming that images with similar structure present similar depth maps. Local Binary Patterns are used to characterize the structure of the color images. The depth maps of those color images with a similar structure to the query image are adaptively combined and filtered to estimate the final depth map. Using public databases, promising results have been obtained outperforming other state-of-the-art algorithms and with a computational cost similar to the most efficient 2D-to-3D algorithms.