Machines de vision et changement de paradigme : une étude comparative de l'imagerie scientifique dans les travaux d'Étienne-Jules Marey et de Marcel Duchamp

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal

Building Beyond Limits : Fantastic Collisions Between Bodies and Machines in French and English Fin-de-Siècle Literature

«Construire hors limite: collisions fantastiques entre corps et machines dans la littérature fin-de-siècle française et anglaise» explore un ensemble de textes qui ont surgi à la fin du dix-neuvième siècle en réponse et en réaction à la fulgurante évolution de l’environnement scientifique et technologique, et qui considèrent la relation entre l’homme et la machine en fantasmant sur la zone grise où ils s’intersectent. Les principaux textes étudiés comprennent L’Ève future de Villiers de l’Isle-Adam, Le Surmâle d’Alfred Jarry, Trilby de George Du Maurier, Le Château des Carpathes de Jules Verne, ainsi qu’une sélection de contes dont nous pouvons qualifier de «contes à appareils», notamment «La Machine à parler» de Marcel Schwob. Utilisant la théorie des systèmes comme base méthodologique, cette dissertation cherche à réinterpréter les textes de la fin du dix-neuvième siècle qui naviguent les limites de l’humain et du mécanique et les surfaces sensibles où ils se touchent et interagissent en les réinscrivant dans un projet plus vaste de construction d’identité qui défie le temps chronologique et les échelles mathématiques. Le lien entre la théorie des systèmes et l’architecture – comme méthode d’organisation d’espace blanc en espace habitable – est exploré dans le but de comprendre la manière dont nous façonnons et interprétons le néant à l’origine de l’identité individuelle, et par association collective, en pratiquant littéralement la schématisation et la construction du corps. Des auteurs tels Villiers et Jarry imaginent la construction du corps comme une entreprise scientifique nécessairement fondée et réalisée avec les matériaux et les technologies disponibles, pour ensuite démanteler cette proposition en condamnant le corps technologique à la destruction. La construction d’une identité amplifiée par la technologie prend donc des proportions prométhéennes perpétuellement redessinées dans des actes cycliques de rasage (destruction) et d’érection (édification), et reflétées dans l’écriture palimpsestique du texte. L’intégrité du corps organique étant mis en question, le noyau même de ce que signifie l’être (dans son sens de verbe infinitif) humain pourrait bien s’avérer, si l’on considère la correspondance entre perte de voix et état pathologique dans les textes de Du Maurier, Verne et Schwob, être une structure des plus précaires, distinctement hors sens (unsound).

Improving sampling, optimization and feature extraction in Boltzmann machines

L’apprentissage supervisé de réseaux hiérarchiques à grande échelle connaît présentement un succès fulgurant. Malgré cette effervescence, l’apprentissage non-supervisé représente toujours, selon plusieurs chercheurs, un élément clé de l’Intelligence Artificielle, où les agents doivent apprendre à partir d’un nombre potentiellement limité de données. Cette thèse s’inscrit dans cette pensée et aborde divers sujets de recherche liés au problème d’estimation de densité par l’entremise des machines de Boltzmann (BM), modèles graphiques probabilistes au coeur de l’apprentissage profond. Nos contributions touchent les domaines de l’échantillonnage, l’estimation de fonctions de partition, l’optimisation ainsi que l’apprentissage de représentations invariantes. Cette thèse débute par l’exposition d’un nouvel algorithme d'échantillonnage adaptatif, qui ajuste (de fa ̧con automatique) la température des chaînes de Markov sous simulation, afin de maintenir une vitesse de convergence élevée tout au long de l’apprentissage. Lorsqu’utilisé dans le contexte de l’apprentissage par maximum de vraisemblance stochastique (SML), notre algorithme engendre une robustesse accrue face à la sélection du taux d’apprentissage, ainsi qu’une meilleure vitesse de convergence. Nos résultats sont présent ́es dans le domaine des BMs, mais la méthode est générale et applicable à l’apprentissage de tout modèle probabiliste exploitant l’échantillonnage par chaînes de Markov. Tandis que le gradient du maximum de vraisemblance peut-être approximé par échantillonnage, l’évaluation de la log-vraisemblance nécessite un estimé de la fonction de partition. Contrairement aux approches traditionnelles qui considèrent un modèle donné comme une boîte noire, nous proposons plutôt d’exploiter la dynamique de l’apprentissage en estimant les changements successifs de log-partition encourus à chaque mise à jour des paramètres. Le problème d’estimation est reformulé comme un problème d’inférence similaire au filtre de Kalman, mais sur un graphe bi-dimensionnel, où les dimensions correspondent aux axes du temps et au paramètre de température. Sur le thème de l’optimisation, nous présentons également un algorithme permettant d’appliquer, de manière efficace, le gradient naturel à des machines de Boltzmann comportant des milliers d’unités. Jusqu’à présent, son adoption était limitée par son haut coût computationel ainsi que sa demande en mémoire. Notre algorithme, Metric-Free Natural Gradient (MFNG), permet d’éviter le calcul explicite de la matrice d’information de Fisher (et son inverse) en exploitant un solveur linéaire combiné à un produit matrice-vecteur efficace. L’algorithme est prometteur: en terme du nombre d’évaluations de fonctions, MFNG converge plus rapidement que SML. Son implémentation demeure malheureusement inefficace en temps de calcul. Ces travaux explorent également les mécanismes sous-jacents à l’apprentissage de représentations invariantes. À cette fin, nous utilisons la famille de machines de Boltzmann restreintes “spike & slab” (ssRBM), que nous modifions afin de pouvoir modéliser des distributions binaires et parcimonieuses. Les variables latentes binaires de la ssRBM peuvent être rendues invariantes à un sous-espace vectoriel, en associant à chacune d’elles, un vecteur de variables latentes continues (dénommées “slabs”). Ceci se traduit par une invariance accrue au niveau de la représentation et un meilleur taux de classification lorsque peu de données étiquetées sont disponibles. Nous terminons cette thèse sur un sujet ambitieux: l’apprentissage de représentations pouvant séparer les facteurs de variations présents dans le signal d’entrée. Nous proposons une solution à base de ssRBM bilinéaire (avec deux groupes de facteurs latents) et formulons le problème comme l’un de “pooling” dans des sous-espaces vectoriels complémentaires.

Non invasive classification system of scoliosis curve types using least-squares support vector machines

Objective To determine scoliosis curve types using non invasive surface acquisition, without prior knowledge from X-ray data. Methods Classification of scoliosis deformities according to curve type is used in the clinical management of scoliotic patients. In this work, we propose a robust system that can determine the scoliosis curve type from non invasive acquisition of the 3D back surface of the patients. The 3D image of the surface of the trunk is divided into patches and local geometric descriptors characterizing the back surface are computed from each patch and constitute the features. We reduce the dimensionality by using principal component analysis and retain 53 components using an overlap criterion combined with the total variance in the observed variables. In this work, a multi-class classifier is built with least-squares support vector machines (LS-SVM). The original LS-SVM formulation was modified by weighting the positive and negative samples differently and a new kernel was designed in order to achieve a robust classifier. The proposed system is validated using data from 165 patients with different scoliosis curve types. The results of our non invasive classification were compared with those obtained by an expert using X-ray images. Results The average rate of successful classification was computed using a leave-one-out cross-validation procedure. The overall accuracy of the system was 95%. As for the correct classification rates per class, we obtained 96%, 84% and 97% for the thoracic, double major and lumbar/thoracolumbar curve types, respectively. Conclusion This study shows that it is possible to find a relationship between the internal deformity and the back surface deformity in scoliosis with machine learning methods. The proposed system uses non invasive surface acquisition, which is safe for the patient as it involves no radiation. Also, the design of a specific kernel improved classification performance.

Concurrent Aggregates (CA): An Object-Oriented Language for Fine-Grained Message-Passing Machines

Fine-grained parallel machines have the potential for very high speed computation. To program massively-concurrent MIMD machines, programmers need tools for managing complexity. These tools should not restrict program concurrency. Concurrent Aggregates (CA) provides multiple-access data abstraction tools, Aggregates, which can be used to implement abstractions with virtually unlimited potential for concurrency. Such tools allow programmers to modularize programs without reducing concurrency. I describe the design, motivation, implementation and evaluation of Concurrent Aggregates. CA has been used to construct a number of application programs. Multi-access data abstractions are found to be useful in constructing highly concurrent programs.

Comparing Support Vector Machines with Gaussian Kernels to Radial Basis Function Classifiers

The Support Vector (SV) machine is a novel type of learning machine, based on statistical learning theory, which contains polynomial classifiers, neural networks, and radial basis function (RBF) networks as special cases. In the RBF case, the SV algorithm automatically determines centers, weights and threshold such as to minimize an upper bound on the expected test error. The present study is devoted to an experimental comparison of these machines with a classical approach, where the centers are determined by $k$--means clustering and the weights are found using error backpropagation. We consider three machines, namely a classical RBF machine, an SV machine with Gaussian kernel, and a hybrid system with the centers determined by the SV method and the weights trained by error backpropagation. Our results show that on the US postal service database of handwritten digits, the SV machine achieves the highest test accuracy, followed by the hybrid approach. The SV approach is thus not only theoretically well--founded, but also superior in a practical application.

Properties of Support Vector Machines

Support Vector Machines (SVMs) perform pattern recognition between two point classes by finding a decision surface determined by certain points of the training set, termed Support Vectors (SV). This surface, which in some feature space of possibly infinite dimension can be regarded as a hyperplane, is obtained from the solution of a problem of quadratic programming that depends on a regularization parameter. In this paper we study some mathematical properties of support vectors and show that the decision surface can be written as the sum of two orthogonal terms, the first depending only on the margin vectors (which are SVs lying on the margin), the second proportional to the regularization parameter. For almost all values of the parameter, this enables us to predict how the decision surface varies for small parameter changes. In the special but important case of feature space of finite dimension m, we also show that there are at most m+1 margin vectors and observe that m+1 SVs are usually sufficient to fully determine the decision surface. For relatively small m this latter result leads to a consistent reduction of the SV number.

Notes on PCA, Regularization, Sparsity and Support Vector Machines

We derive a new representation for a function as a linear combination of local correlation kernels at optimal sparse locations and discuss its relation to PCA, regularization, sparsity principles and Support Vector Machines. We first review previous results for the approximation of a function from discrete data (Girosi, 1998) in the context of Vapnik"s feature space and dual representation (Vapnik, 1995). We apply them to show 1) that a standard regularization functional with a stabilizer defined in terms of the correlation function induces a regression function in the span of the feature space of classical Principal Components and 2) that there exist a dual representations of the regression function in terms of a regularization network with a kernel equal to a generalized correlation function. We then describe the main observation of the paper: the dual representation in terms of the correlation function can be sparsified using the Support Vector Machines (Vapnik, 1982) technique and this operation is equivalent to sparsify a large dictionary of basis functions adapted to the task, using a variation of Basis Pursuit De-Noising (Chen, Donoho and Saunders, 1995; see also related work by Donahue and Geiger, 1994; Olshausen and Field, 1995; Lewicki and Sejnowski, 1998). In addition to extending the close relations between regularization, Support Vector Machines and sparsity, our work also illuminates and formalizes the LFA concept of Penev and Atick (1996). We discuss the relation between our results, which are about regression, and the different problem of pattern classification.

From Regression to Classification in Support Vector Machines

We study the relation between support vector machines (SVMs) for regression (SVMR) and SVM for classification (SVMC). We show that for a given SVMC solution there exists a SVMR solution which is equivalent for a certain choice of the parameters. In particular our result is that for $epsilon$ sufficiently close to one, the optimal hyperplane and threshold for the SVMC problem with regularization parameter C_c are equal to (1-epsilon)^{- 1} times the optimal hyperplane and threshold for SVMR with regularization parameter C_r = (1-epsilon)C_c. A direct consequence of this result is that SVMC can be seen as a special case of SVMR.

A Unified Framework for Regularization Networks and Support Vector Machines

Regularization Networks and Support Vector Machines are techniques for solving certain problems of learning from examples -- in particular the regression problem of approximating a multivariate function from sparse data. We present both formulations in a unified framework, namely in the context of Vapnik's theory of statistical learning which provides a general foundation for the learning problem, combining functional analysis and statistics.

An Equivalence Between Sparse Approximation and Support Vector Machines

In the first part of this paper we show a similarity between the principle of Structural Risk Minimization Principle (SRM) (Vapnik, 1982) and the idea of Sparse Approximation, as defined in (Chen, Donoho and Saunders, 1995) and Olshausen and Field (1996). Then we focus on two specific (approximate) implementations of SRM and Sparse Approximation, which have been used to solve the problem of function approximation. For SRM we consider the Support Vector Machine technique proposed by V. Vapnik and his team at AT&T Bell Labs, and for Sparse Approximation we consider a modification of the Basis Pursuit De-Noising algorithm proposed by Chen, Donoho and Saunders (1995). We show that, under certain conditions, these two techniques are equivalent: they give the same solution and they require the solution of the same quadratic programming problem.

Support Vector Machines: Training and Applications

The Support Vector Machine (SVM) is a new and very promising classification technique developed by Vapnik and his group at AT&T Bell Labs. This new learning algorithm can be seen as an alternative training technique for Polynomial, Radial Basis Function and Multi-Layer Perceptron classifiers. An interesting property of this approach is that it is an approximate implementation of the Structural Risk Minimization (SRM) induction principle. The derivation of Support Vector Machines, its relationship with SRM, and its geometrical insight, are discussed in this paper. Training a SVM is equivalent to solve a quadratic programming problem with linear and box constraints in a number of variables equal to the number of data points. When the number of data points exceeds few thousands the problem is very challenging, because the quadratic form is completely dense, so the memory needed to store the problem grows with the square of the number of data points. Therefore, training problems arising in some real applications with large data sets are impossible to load into memory, and cannot be solved using standard non-linear constrained optimization algorithms. We present a decomposition algorithm that can be used to train SVM's over large data sets. The main idea behind the decomposition is the iterative solution of sub-problems and the evaluation of, and also establish the stopping criteria for the algorithm. We present previous approaches, as well as results and important details of our implementation of the algorithm using a second-order variant of the Reduced Gradient Method as the solver of the sub-problems. As an application of SVM's, we present preliminary results we obtained applying SVM to the problem of detecting frontal human faces in real images.

A Note on Support Vector Machines Degeneracy

When training Support Vector Machines (SVMs) over non-separable data sets, one sets the threshold $b$ using any dual cost coefficient that is strictly between the bounds of $0$ and $C$. We show that there exist SVM training problems with dual optimal solutions with all coefficients at bounds, but that all such problems are degenerate in the sense that the "optimal separating hyperplane" is given by ${f w} = {f 0}$, and the resulting (degenerate) SVM will classify all future points identically (to the class that supplies more training data). We also derive necessary and sufficient conditions on the input data for this to occur. Finally, we show that an SVM training problem can always be made degenerate by the addition of a single data point belonging to a certain unboundedspolyhedron, which we characterize in terms of its extreme points and rays.

The Age of Social Machines

Many of the most successful and important systems that impact our lives combine humans, data, and algorithms at Web Scale. These social machines are amalgamations of human and machine intelligence. This seminar will provide an update on SOCIAM, a five year EPSRC Programme Grant that seeks to gain a better understanding of social machines; how they are observed and constituted, how they can be designed and their fate determined. We will review how social machines can be of value to society, organisations and individuals. We will consider the challenges they present to our various disciplines.

Hierarchical Prediction Machines and Big Data Analytics

An emerging consensus in cognitive science views the biological brain as a hierarchically-organized predictive processing system. This is a system in which higher-order regions are continuously attempting to predict the activity of lower-order regions at a variety of (increasingly abstract) spatial and temporal scales. The brain is thus revealed as a hierarchical prediction machine that is constantly engaged in the effort to predict the flow of information originating from the sensory surfaces. Such a view seems to afford a great deal of explanatory leverage when it comes to a broad swathe of seemingly disparate psychological phenomena (e.g., learning, memory, perception, action, emotion, planning, reason, imagination, and conscious experience). In the most positive case, the predictive processing story seems to provide our first glimpse at what a unified (computationally-tractable and neurobiological plausible) account of human psychology might look like. This obviously marks out one reason why such models should be the focus of current empirical and theoretical attention. Another reason, however, is rooted in the potential of such models to advance the current state-of-the-art in machine intelligence and machine learning. Interestingly, the vision of the brain as a hierarchical prediction machine is one that establishes contact with work that goes under the heading of 'deep learning'. Deep learning systems thus often attempt to make use of predictive processing schemes and (increasingly abstract) generative models as a means of supporting the analysis of large data sets. But are such computational systems sufficient (by themselves) to provide a route to general human-level analytic capabilities? I will argue that they are not and that closer attention to a broader range of forces and factors (many of which are not confined to the neural realm) may be required to understand what it is that gives human cognition its distinctive (and largely unique) flavour. The vision that emerges is one of 'homomimetic deep learning systems', systems that situate a hierarchically-organized predictive processing core within a larger nexus of developmental, behavioural, symbolic, technological and social influences. Relative to that vision, I suggest that we should see the Web as a form of 'cognitive ecology', one that is as much involved with the transformation of machine intelligence as it is with the progressive reshaping of our own cognitive capabilities.