In this thesis, our aim is to improve deep auto-encoders, an important topic in the deep learning area, which has shown connections to latent feature discovery models in the literature. Our model is inspired by robust principal component analysis, and we build an outlier filter on the top of basic deep auto-encoders. By adding this filter, we can split the input data X into two parts X=L+S, where the L could be better reconstructed by a deep auto-encoder and the S contains the anomalous parts of the original data X. Filtering out the anomalies increases the robustness of the standard auto-encoder, and thus we name our model ``Robust Auto-encoder'. We also propose a novel solver for the robust auto-encoder which alternatively optimizes the reconstruction cost of the deep auto-encoder and the sparsity of outlier filter in pursuit of finding the optimal solution. This solver is inspired by the Alternating Direction Method of Multipliers, Back-propagation and the Alternating Projection method, and we demonstrate the convergence properties of this algorithm and its superior performance in standard image recognition tasks. Last but not least, we apply our model to multiple domains, especially, the cyber-data analysis, where deep models are seldom currently used.

In the ever-changing landscape of the fishing industry, demands for automating specific processes are increasing substantially. Predicting future events eliminates much of the existing communication latency between fishing vessels and their customers and makes real-time analysis of onboard catch possible for the fishing industry. Further, machine learning models, may reduce the number of human resources necessary for the numerous processes that may be automated. In this document, we focus on weight estimation of three different species of fish. Namely, we want to estimate the fish weight given its specie through datadriven techniques. Due to the high complexity of image data, the overhead expenses of collecting images at sea, and the complexities of fish features, we consider a dimensionality reduction on the inputs to reduce the curse of dimensionality and increase interpretability. We will study the viability of modeling fish weights from lower-dimensional feature vectors and the conjunction of lower-dimensional feature vectors and algorithmically obtained features. We found that modeling the residuals with latent representations of a simple power model fitted on length features resulted in a significant difference in the weight estimates for two types of fish and a decrease in Root Mean Squared Error (rMSE) and Mean Absolute Percentage Error (MAPE) scores in favour of the estimations utilizing latent representations. / I fiskeindustrins ständigt föränderliga landskap ökar kraven på att automatisera specifika processer väsentligt. Att förutsäga framtida händelser eliminerar mycket av den befintliga kommunikationsfördröjningen mellan fiskefartyg och deras kunder och möjliggör analys i realtid av ombordfångst för fiskeindustrin. Vidare kan det minska antalet personalresurser som krävs för de många processer som kan automatiseras. I detta dokument studerar vi två olika beslutsproblem relaterade till att sortera fisk av tre olika arter. Vi vill nämligen bestämma fiskvikten och dess art genom datadrivna tekniker. På grund av bilddatas höga komplexitet, de allmänna kostnaderna för att samla bilder till sjöss och komplexiteten hos fiskegenskaper, anser vi att en dimensionalitetsminskning av särdragen minskar problemet relaterat till dimensionsexplosion och ökar tolkbarheten. Vi kommer att studera lämpligheten av modellering av fiskvikter och arter från lägre dimensionella särdragsvektorer samt kombinationen av dessa med algoritmiskt erhållna funktioner. Vi fann att modellering av residual med latenta representationer av en enkel potensfunktionsmodell som är anpassad till fisklängder resulterade i en signifikant skillnad i viktuppskattningarna för två typer av fisk och en minskning av rMSE och MAPE poäng.

Ingeniero Civil Mecánico / Dentro del campo de la ingeniería mecánica, una de las áreas que más crecimiento ha mostrado en los últimos años es la de la gestión de activos físicos y confiabilidad. Junto con la capacidad de construir máquinas y sistemas más complejos, el problema de la detección temprana de fallas en elementos mecánicos se vuelve de suma importancia. Al mismo tiempo, el incremento en la disponibilidad de tecnología sensitoria ha dado a los ingenieros la capacidad de medir una gran cantidad de variables operacionales, como por ejemplo presión, temperatura o emisiones acústicas, a frecuencias de muestreo altísimas. Es ese aspecto, se vuelve un desafío en sí mismo el poder procesar esa cantidad de datos de una manera eficiente, con tal de extrar información útil a partir de ellos. Una metodología para enfrentar este problema es el desarrollo de técnicas de reducción de dimensionalidad, las cuales, si son implementadas de forma correcta, pueden generar una mejor representación de los datos con el fin de mejorar el diagnóstico posterior de los modos de falla presentes. La motivación principal de este trabajo de título es la necesidad de desarrollar modelos confiables para el diagnóstico de modos de falla en elementos mecánicos utilizando técnicas de Aprendizaje de Máquinas. Estos modelos pueden resultar en grandes beneficios para los sectores industriales, tanto en términos de ahorros monetarios como seguridad operacional. El principal objetivo de esta tesis es desarrollar modelos para el diagnóstico de fallas en elementos mecánicos basados en una reducción de dimensionalidad usando un Auto Encoder Variacional (VAE), y luego evaluar y comparar los resultados obtenidos con un modelo similar que usa Análisis de Componentes Principales (PCA) como método de reducción de dimensionalidad y un tercer modelo que no genera una reducción. La metodología usada para este trabajo consiste principalmente de cinco etapas. Primero, una revisión del estado del arte respecto a metodologías existentes para el diagnóstico de fallas es desarrollada. Luego, la adquisición y preprocesamiento de datos operacionales que serán utilizados para entrenar y evaluar los modelos desarrollados. Tercero, el modelo que usa PCA y el modelo que no realiza reducción de dimensionalidad es implementado. Cuarto, el modelo que utiliza VAE es desarrollado e implementado. Por último, el modelo que usa VAE es comparado con los otros dos modelos para extraer conclusiones sobre su aplicabilidad. La principal conclusión de este trabajo es que el modelo que utiliza VAE es mejor en el diagnóstico de modos de falla que el que utiliza PCA para situaciones donde la cantidad de datos etiquetados es escasa, o para los casos cuando una reducción de dimensionalidad muy drástica es requerida. Tambien, el modelo que utiliza VAE casi siempre presenta mejores resultados que el modelo que no genera reducción en los datos, mostrando la importancia de reducir la dimensionalidad de los datos previo a una operación de diagnóstico o clasificación.

[pt] Esta dissertação propõe e avalia estratégias baseadas nos Stacked Supervised Auto-encoders (SSAE) para representação de imagens faciais em aplicações de vídeo vigilância. O estudo foca na identificação de faces a partir de uma amostra por pessoa na galeria (single sample per person - SSPP). Variações em termos de pose, expressão facial, iluminação e oclusão são abordadas de duas formas. Primeiro, o SSAE extrai atributos das imagens de faces que são robustos contra tais variações. Segundo, exploram-se as múltiplas amostras que podem ser coletadas nas sequências de vídeo de uma pessoa (multiple samples per person probe - MSPPP). Os métodos propostos foram avaliados e comparados usando os bancos de vídeos Honda/UCSD e VIDTIMIT. Adicionalmente, foi estudada a influência de parâmetros relacionados com a arquitetura do SSAE utilizando o banco de imagens estáticas Extended Yale B. Os resultados demonstraram que as estratégias que exploram as MSPPP em combinação com o SSAE podem superar o desempenho de outros métodos SSPP, como os Padrões Binários Locais (LBP), para reconhecimento de faces em vídeos. / [en] This work proposes and evaluates strategies based on Stacked Supervised Auto-encoders (SSAE) for face representation in video surveillance applications. The study focuses on the identification task with a single sample per person (SSPP) in the gallery. Variations in terms of pose, facial expression, illumination and occlusion are approached in two ways. First, the SSAE extracts features from face images, which are robust to such variations. Second, multiple samples per persons probes (MSPPP) that can be extracted from video sequences are exploited to improve recognition accuracy. The proposed methods were compared upon Honda/UCSD and VIDTIMIT video datasets. Additionally, the influence of the parameters related to SSAE architecture was studied using the Extended Yale B dataset. The experimental results demonstrated that strategies combining SSAE and MSPPP are able to outperform other SSPP methods, such as local binary patterns, in face recognition from video.

Focal stacking is a technique that allows us to create images with a large depth of field, where everything in the scene is sharp and clear. However, creating such images is not easy, as it requires taking multiple pictures at different focus settings and then blending them together. In this paper, we present a novel approach to blending a focal stack using a special type of autoencoder, which is a neural network that can learn to compress and reconstruct data. Our autoencoder consists of several parts, each of which processes one input image and passes its information to the final part, which fuses them into one output image. Unlike other methods, our approach is capable of inpainting and denoising resulting in sharp, clean all-in-focus images. Our approach does not require any prior training or a large dataset, which makes it fast and effective. We evaluate our method on various kinds of images and compare it with other widely used methods. We demonstrate that our method can produce superior focal stacked images with higher accuracy and quality. This paper reveals a new and promising way of using a neural network to aid in microphotography, microscopy, and visual computing, by enhancing the quality of focal stacked images.

Due to automation, the world is changing at a rapid pace. Autonomous agents have become more common over the last several years and, as a result, have created a need for improved software to back them up. The most important aspect of this greater software is path prediction, as robots need to be able to decide where to move in the future. In order to accomplish this, a robot must know how to avoid humans, putting frame prediction at the core of many modern day solutions. A popular way to solve this complex problem of frame prediction is Auto Encoder LSTMs. Though there are many implementations of this, at its core, it is a neural network comprised of a series of time sensitive processing blocks that shrink and then grow the data’s dimensions to make a prediction. The idea of using Auto Encoder styled networks to do frame prediction has also been adapted by others to make Temporal Encoders. These neural networks work much like traditional Auto Encoders, in which the data is reduced then expanded back up. These networks attempt to tease out a series of frames, including a predictive frame of the future. The problem with many of these networks is that they take an immense amount of computation power, and time to get them performing at an acceptable level. This thesis presents possible ways of pre-processing input frames to these networks in order to gain performance, in the best case seeing a 360x improvement in accuracy compared to the original models. This thesis also extends the work done with Temporal Encoders to create more precise prediction models, which showed consistent improvements of at least 50% for some metrics. All of the generated models were compared using a simulated data set collected from recordings of ground level viewpoints from Cities: Skylines. These predicted frames were then analyzed using a common perceptual distance metric, that is, Minkowski distance, as well as a custom metric that tracked distinct areas in frames. All of the following was run on a constrained system in order to see the effects of the changes as they pertain to systems with limited hardware access.

No description available.

Understanding how users interact with a company's service is essential for data-driven businesses that want to better cater to their users and improve their offering. By using a generative machine learning approach it is possible to model user behaviour and generate new data to simulate or recognize and explain typical usage patterns. In this work we introduce an approach for modelling users' interaction behaviour at scale in a client-service model. We propose a novel representation of multivariate time-series data as time pictures that express temporal correlations through spatial organization. This representation shares two key properties that convolutional networks have been built to exploit and allows us to develop an approach based on deep generative models that use convolutional networks as backbone. In introducing this approach of feature learning for time-series data, we expand the application of convolutional neural networks in the multivariate time-series domain, and specifically user interaction data. We adopt a variational approach inspired by the β-VAE framework in order to learn hidden factors that define different user behaviour patterns. We explore different values for the regularization parameter β and show that it is possible to construct a model that learns a latent representation of identifiable and different user behaviours. We show on real-world data that the model generates realistic samples, that capture the true population-level statistics of the interaction behaviour data, learns different user behaviours, and provides accurate imputations of missing data. / Förståelse för hur användare interagerar med ett företags tjänst är essentiell för data-drivna affärsverksamheter med ambitioner om att bättre tillgodose dess användare och att förbättra deras utbud. Generativ maskininlärning möjliggör modellering av användarbeteende och genererande av ny data i syfte att simulera eller identifiera och förklara typiska användarmönster. I detta arbete introducerar vi ett tillvägagångssätt för storskalig modellering av användarinteraktion i en klientservice-modell. Vi föreslår en ny representation av multivariat tidsseriedata i form av tidsbilder vilka representerar temporala korrelationer via spatial organisering. Denna representation delar två nyckelegenskaper som faltningsnätverk har utvecklats för att exploatera, vilket tillåter oss att utveckla ett tillvägagångssätt baserat på på djupa generativa modeller som bygger på faltningsnätverk. Genom att introducera detta tillvägagångssätt för tidsseriedata expanderar vi applicering av faltningsnätverk inom domänen för multivariat tidsserie, specifikt för användarinteraktionsdata. Vi använder ett tillvägagångssätt inspirerat av ramverket β-VAE i syfte att lära modellen gömda faktorer som definierar olika användarmönster. Vi utforskar olika värden för regulariseringsparametern β och visar att det är möjligt att konstruera en modell som lär sig en latent representation av identifierbara och multipla användarbeteenden. Vi visar med verklig data att modellen genererar realistiska exempel vilka i sin tur fångar statistiken på populationsnivå hos användarinteraktionsdatan, samt lär olika användarbeteenden och bidrar med precisa imputationer av saknad data.

An issue authorities and banks face is fraud related to payments and transactions where huge monetary losses occur to a party or where money laundering schemes are carried out. Previous work in the field of machine learning for fraud detection has addressed the issue as a supervised learning problem. In this thesis, we propose a model which can be used in a fraud detection system with transactions and payments that are unlabeled. The proposed modelis a Long Short-term Memory in an auto-encoder decoder network (LSTMAED)which is trained and tested on transformed data. The data is transformed by reducing it to Principal Components and clustering it with K-means. The model is trained to reconstruct the sequence with high accuracy. Our results indicate that the LSTM-AED performs better than a random sequence generating process in learning and reconstructing a sequence of payments. We also found that huge a loss of information occurs in the pre-processing stages. / Obehöriga transaktioner och bedrägerier i betalningar kan leda till stora ekonomiska förluster för banker och myndigheter. Inom maskininlärning har detta problem tidigare hanterats med hjälp av klassifierare via supervised learning. I detta examensarbete föreslår vi en modell som kan användas i ett system för att upptäcka bedrägerier. Modellen appliceras på omärkt data med många olika variabler. Modellen som används är en Long Short-term memory i en auto-encoder decoder nätverk. Datan transformeras med PCA och klustras med K-means. Modellen tränas till att rekonstruera en sekvens av betalningar med hög noggrannhet. Vår resultat visar att LSTM-AED presterar bättre än en modell som endast gissar nästa punkt i sekvensen. Resultatet visar också att mycket information i datan går förlorad när den förbehandlas och transformeras.

In this master’s thesis we explore the idea of using artificial intelligence in the product design process and seek to develop a conceptual framework for how it can be incorporated to make user customized products more accessible and affordable for everyone. We show how generative deep learning models such as Variational Auto Encoders and Generative Adversarial Networks can be implemented to generate design variations of windows and clarify the general implementation process along with insights from recent research in the field. The proposed framework consists of three parts: (1) A morphological matrix connecting several identified possibilities of implementation to specific parts of the product design process. (2) A general step-by-step process on how to incorporate generative deep learning. (3) A description of common challenges, strategies andsolutions related to the implementation process. Together with the framework we also provide a system for automatic gathering and cleaning of image data as well as a dataset containing 4564 images of windows in a front view perspective.