Detecting social patterns within 20th century documentary photos: a deep learning based approach

The job of a historian is to understand what happened in the past, resorting in many cases to written documents as a firsthand source of information. Text, however, does not amount to the only source of knowledge. Pictorial representations, in fact, have also accompanied the main events of the historical timeline. In particular, the opportunity of visually representing circumstances has bloomed since the invention of photography, with the possibility of capturing in real-time the occurrence of a specific events. Thanks to the widespread use of digital technologies (e.g. smartphones and digital cameras), networking capabilities and consequent availability of multimedia content, the academic and industrial research communities have developed artificial intelligence (AI) paradigms with the aim of inferring, transferring and creating new layers of information from images, videos, etc. Now, while AI communities are devoting much of their attention to analyze digital images, from an historical research standpoint more interesting results may be obtained analyzing analog images representing the pre-digital era. Within the aforementioned scenario, the aim of this work is to analyze a collection of analog documentary photographs, building upon state-of-the-art deep learning techniques. In particular, the analysis carried out in this thesis aims at producing two following results: (a) produce the date of an image, and, (b) recognizing its background socio-cultural context,as defined by a group of historical-sociological researchers. Given these premises, the contribution of this work amounts to: (i) the introduction of an historical dataset including images of “Family Album” among all the twentieth century, (ii) the introduction of a new classification task regarding the identification of the socio-cultural context of an image, (iii) the exploitation of different deep learning architectures to perform the image dating and the image socio-cultural context classification.

Smart sampling approaches for Decision-Focused Learning

Many real-word decision- making problems are defined based on forecast parameters: for example, one may plan an urban route by relying on traffic predictions. In these cases, the conventional approach consists in training a predictor and then solving an optimization problem. This may be problematic since mistakes made by the predictor may trick the optimizer into taking dramatically wrong decisions. Recently, the field of Decision-Focused Learning overcomes this limitation by merging the two stages at training time, so that predictions are rewarded and penalized based on their outcome in the optimization problem. There are however still significant challenges toward a widespread adoption of the method, mostly related to the limitation in terms of generality and scalability. One possible solution for dealing with the second problem is introducing a caching-based approach, to speed up the training process. This project aims to investigate these techniques, in order to reduce even more, the solver calls. For each considered method, we designed a particular smart sampling approach, based on their characteristics. In the case of the SPO method, we ended up discovering that it is only necessary to initialize the cache with only several solutions; those needed to filter the elements that we still need to properly learn. For the Blackbox method, we designed a smart sampling approach, based on inferred solutions.

Apprentissage d'espaces sémantiques

Dans cette dissertation, nous présentons plusieurs techniques d’apprentissage d’espaces sémantiques pour plusieurs domaines, par exemple des mots et des images, mais aussi à l’intersection de différents domaines. Un espace de représentation est appelé sémantique si des entités jugées similaires par un être humain, ont leur similarité préservée dans cet espace. La première publication présente un enchaînement de méthodes d’apprentissage incluant plusieurs techniques d’apprentissage non supervisé qui nous a permis de remporter la compétition “Unsupervised and Transfer Learning Challenge” en 2011. Le deuxième article présente une manière d’extraire de l’information à partir d’un contexte structuré (177 détecteurs d’objets à différentes positions et échelles). On montrera que l’utilisation de la structure des données combinée à un apprentissage non supervisé permet de réduire la dimensionnalité de 97% tout en améliorant les performances de reconnaissance de scènes de +5% à +11% selon l’ensemble de données. Dans le troisième travail, on s’intéresse à la structure apprise par les réseaux de neurones profonds utilisés dans les deux précédentes publications. Plusieurs hypothèses sont présentées et testées expérimentalement montrant que l’espace appris a de meilleures propriétés de mixage (facilitant l’exploration de différentes classes durant le processus d’échantillonnage). Pour la quatrième publication, on s’intéresse à résoudre un problème d’analyse syntaxique et sémantique avec des réseaux de neurones récurrents appris sur des fenêtres de contexte de mots. Dans notre cinquième travail, nous proposons une façon d’effectuer de la recherche d’image ”augmentée” en apprenant un espace sémantique joint où une recherche d’image contenant un objet retournerait aussi des images des parties de l’objet, par exemple une recherche retournant des images de ”voiture” retournerait aussi des images de ”pare-brises”, ”coffres”, ”roues” en plus des images initiales.

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.

Inference From Aging Information

For many learning tasks the duration of the data collection can be greater than the time scale for changes of the underlying data distribution. The question we ask is how to include the information that data are aging. Ad hoc methods to achieve this include the use of validity windows that prevent the learning machine from making inferences based on old data. This introduces the problem of how to define the size of validity windows. In this brief, a new adaptive Bayesian inspired algorithm is presented for learning drifting concepts. It uses the analogy of validity windows in an adaptive Bayesian way to incorporate changes in the data distribution over time. We apply a theoretical approach based on information geometry to the classification problem and measure its performance in simulations. The uncertainty about the appropriate size of the memory windows is dealt with in a Bayesian manner by integrating over the distribution of the adaptive window size. Thus, the posterior distribution of the weights may develop algebraic tails. The learning algorithm results from tracking the mean and variance of the posterior distribution of the weights. It was found that the algebraic tails of this posterior distribution give the learning algorithm the ability to cope with an evolving environment by permitting the escape from local traps.

Predição de promotores de Bacillus subtilis usando técnicas de aprendizado de máquina

One of the most important goals of bioinformatics is the ability to identify genes in uncharacterized DNA sequences on world wide database. Gene expression on prokaryotes initiates when the RNA-polymerase enzyme interacts with DNA regions called promoters. In these regions are located the main regulatory elements of the transcription process. Despite the improvement of in vitro techniques for molecular biology analysis, characterizing and identifying a great number of promoters on a genome is a complex task. Nevertheless, the main drawback is the absence of a large set of promoters to identify conserved patterns among the species. Hence, a in silico method to predict them on any species is a challenge. Improved promoter prediction methods can be one step towards developing more reliable ab initio gene prediction methods. In this work, we present an empirical comparison of Machine Learning (ML) techniques such as Na¨ýve Bayes, Decision Trees, Support Vector Machines and Neural Networks, Voted Perceptron, PART, k-NN and and ensemble approaches (Bagging and Boosting) to the task of predicting Bacillus subtilis. In order to do so, we first built two data set of promoter and nonpromoter sequences for B. subtilis and a hybrid one. In order to evaluate of ML methods a cross-validation procedure is applied. Good results were obtained with methods of ML like SVM and Naïve Bayes using B. subtilis. However, we have not reached good results on hybrid database

Segmentação interativa de imagens utilizando competição e cooperação entre partículas

Pós-graduação em Ciência da Computação - IBILCE

Apprendimento di alberi di decisione

In questa tesi viene effettuata una panoramica sull'apprendimento di alberi di decisione (un particolare metodo di apprendimento supervisionato) e sulle varie tecniche per poterne migliorare le prestazioni. Vengono poi descritti alcuni algoritmi e ne vengono confrontate le prestazioni.

Multiclassificador inteligente de falhas no domínio do tempo em motores de indução trifásicos alimentados por inversores de frequência

Os motores de indução desempenham um importante papel na indústria, fato este que destaca a importância do correto diagnóstico e classificação de falhas ainda em fase inicial de sua evolução, possibilitando aumento na produtividade e, principalmente, eliminando graves danos aos processos e às máquinas. Assim, a proposta desta tese consiste em apresentar um multiclassificador inteligente para o diagnóstico de motor sem defeitos, falhas de curto-circuito nos enrolamentos do estator, falhas de rotor e falhas de rolamentos em motores de indução trifásicos acionados por diferentes modelos de inversores de frequência por meio da análise das amplitudes dos sinais de corrente de estator no domínio do tempo. Para avaliar a precisão de classificação frente aos diversos níveis de severidade das falhas, foram comparados os desempenhos de quatro técnicas distintas de aprendizado de máquina; a saber: (i) Rede Fuzzy Artmap, (ii) Rede Perceptron Multicamadas, (iii) Máquina de Vetores de Suporte e (iv) k-Vizinhos-Próximos. Resultados experimentais obtidos a partir de 13.574 ensaios experimentais são apresentados para validar o estudo considerando uma ampla faixa de frequências de operação, bem como regimes de conjugado de carga em 5 motores diferentes.

Aplicaciones de procesamiento de lenguaje natural

El campo de procesamiento de lenguaje natural (PLN), ha tenido un gran crecimiento en los últimos años; sus áreas de investigación incluyen: recuperación y extracción de información, minería de datos, traducción automática, sistemas de búsquedas de respuestas, generación de resúmenes automáticos, análisis de sentimientos, entre otras. En este artículo se presentan conceptos y algunas herramientas con el fin de contribuir al entendimiento del procesamiento de texto con técnicas de PLN, con el propósito de extraer información relevante que pueda ser usada en un gran rango de aplicaciones. Se pueden desarrollar clasificadores automáticos que permitan categorizar documentos y recomendar etiquetas; estos clasificadores deben ser independientes de la plataforma, fácilmente personalizables para poder ser integrados en diferentes proyectos y que sean capaces de aprender a partir de ejemplos. En el presente artículo se introducen estos algoritmos de clasificación, se analizan algunas herramientas de código abierto disponibles actualmente para llevar a cabo estas tareas y se comparan diversas implementaciones utilizando la métrica F en la evaluación de los clasificadores.

Induction by a Hilbert hypercube representation

This thesis describes a novel connectionist machine utilizing induction by a Hilbert hypercube representation. This representation offers a number of distinct advantages which are described. We construct a theoretical and practical learning machine which lies in an area of overlap between three disciplines - neural nets, machine learning and knowledge acquisition - hence it is refered to as a "coalesced" machine. To this unifying aspect is added the various advantages of its orthogonal lattice structure as against less structured nets. We discuss the case for such a fundamental and low level empirical learning tool and the assumptions behind the machine are clearly outlined. Our theory of an orthogonal lattice structure the Hilbert hypercube of an n-dimensional space using a complemented distributed lattice as a basis for supervised learning is derived from first principles on clearly laid out scientific principles. The resulting "subhypercube theory" was implemented in a development machine which was then used to test the theoretical predictions again under strict scientific guidelines. The scope, advantages and limitations of this machine were tested in a series of experiments. Novel and seminal properties of the machine include: the "metrical", deterministic and global nature of its search; complete convergence invariably producing minimum polynomial solutions for both disjuncts and conjuncts even with moderate levels of noise present; a learning engine which is mathematically analysable in depth based upon the "complexity range" of the function concerned; a strong bias towards the simplest possible globally (rather than locally) derived "balanced" explanation of the data; the ability to cope with variables in the network; and new ways of reducing the exponential explosion. Performance issues were addressed and comparative studies with other learning machines indicates that our novel approach has definite value and should be further researched.

Caractérisation de minéraux indicateurs par imagerie hyperspectrale et traitement de l'image dans l'infrarouge proche et l'infrarouge lointain

L’imagerie hyperspectrale (HSI) fournit de l’information spatiale et spectrale concernant l’émissivité de la surface des matériaux, ce qui peut être utilisée pour l’identification des minéraux. Pour cela, un matériel de référence ou endmember, qui en minéralogie est la forme la plus pure d’un minéral, est nécessaire. L’objectif principal de ce projet est l’identification des minéraux par imagerie hyperspectrale. Les informations de l’imagerie hyperspectrale ont été enregistrées à partir de l’énergie réfléchie de la surface du minéral. L’énergie solaire est la source d’énergie dans l’imagerie hyperspectrale de télédétection, alors qu’un élément chauffant est la source d’énergie utilisée dans les expériences de laboratoire. Dans la première étape de ce travail, les signatures spectrales des minéraux purs sont obtenues avec la caméra hyperspectrale, qui mesure le rayonnement réfléchi par la surface des minéraux. Dans ce projet, deux séries d’expériences ont été menées dans différentes plages de longueurs d’onde (0,4 à 1 µm et 7,7 à 11,8 µm). Dans la deuxième partie de ce projet, les signatures spectrales obtenues des échantillons individuels sont comparées avec des signatures spectrales de la bibliothèque hyperspectrale de l’ASTER. Dans la troisième partie, trois méthodes différentes de classification hyperspectrale sont considérées pour la classification. Spectral Angle Mapper (SAM), Spectral Information Divergence (SID), et Intercorrélation normalisée (NCC). Enfin, un système d’apprentissage automatique, Extreme Learning Machine (ELM), est utilisé pour identifier les minéraux. Deux types d’échantillons ont été utilisés dans ce projet. Le système d’ELM est divisé en deux parties, la phase d’entraînement et la phase de test du système. Dans la phase d’entraînement, la signature d’un seul échantillon minéral est entrée dans le système, et dans la phase du test, les signatures spectrales des différents minéraux, qui sont entrées dans la phase d’entraînement, sont comparées par rapport à des échantillons de minéraux mixtes afin de les identifier.

Um estudo de descritores e classificadores para reconhecimento de bagas de café.

RESUMO - Métodos de reconhecimento de frutos baseados na utilização de diferentes descritores e classificadores foram estudados. Foi utilizada uma base de dados de 3.393 imagens de café e não-café anteriormente criada e rotulada manualmente. Testes quantitativos demonstraram a identificação de bagas com 93% de precisão e 77% de cobertura utilizando descritores HoG adicionados a mediana dos componentes de cor do formato La*b*, aliados ao classificador Gradient Boosting. Esses resultados melhoram o método anteriormente proposto por Santos (2015), e demonstram a possibilidade de evolução de métodos que podem ser aplicados em metodologias de agricultura de precisão, monitoramento e predição de safra.

Exact Combinatorial Optimization with Graph Convolutional Neural Networks

Combinatorial optimization problems are typically tackled by the branch-and-bound paradigm. We propose to learn a variable selection policy for branch-and-bound in mixed-integer linear programming, by imitation learning on a diversified variant of the strong branching expert rule. We encode states as bipartite graphs and parameterize the policy as a graph convolutional neural network. Experiments on a series of synthetic problems demonstrate that our approach produces policies that can improve upon expert-designed branching rules on large problems, and generalize to instances significantly larger than seen during training.

An Embedded In-Cabin Lightweight High-Performance 3D Gaze Estimation System

Gaze estimation has gained interest in recent years for being an important cue to obtain information about the internal cognitive state of humans. Regardless of whether it is the 3D gaze vector or the point of gaze (PoG), gaze estimation has been applied in various fields, such as: human robot interaction, augmented reality, medicine, aviation and automotive. In the latter field, as part of Advanced Driver-Assistance Systems (ADAS), it allows the development of cutting-edge systems capable of mitigating road accidents by monitoring driver distraction. Gaze estimation can be also used to enhance the driving experience, for instance, autonomous driving. It also can improve comfort with augmented reality components capable of being commanded by the driver's eyes. Although, several high-performance real-time inference works already exist, just a few are capable of working with only a RGB camera on computationally constrained devices, such as a microcontroller. This work aims to develop a low-cost, efficient and high-performance embedded system capable of estimating the driver's gaze using deep learning and a RGB camera. The proposed system has achieved near-SOTA performances with about 90% less memory footprint. The capabilities to generalize in unseen environments have been evaluated through a live demonstration, where high performance and near real-time inference were obtained using a webcam and a Raspberry Pi4.