Laser absorption spectroscopy has been a valuable technique for sensitive, non-intrusive, in-situ detection of gaseous and liquid phase target species. The infrared spectral region is specifically attractive as it provides opportunities for selective sensing of a multitude of species in various applications. This thesis explores techniques for interference-free sensing in the infrared region for environmental, combustion, and petrochemical applications. A mid-infrared laser-based sensor was designed to detect trace amounts of benzene using off-axis cavity-enhanced absorption spectroscopy and a multidimensional linear regression algorithm. This sensor achieved unprecedented detection limits, making it ideal for environmental and occupational pollution monitoring. Moreover, wavelength tuning and deep neural networks were employed to differentiate between the broadband similar-shaped absorbance spectra of benzene, toluene, ethylbenzene, and xylene isomers. Benzene sensing was enhanced by recent advancement in semiconductor laser technology, which enabled access to the long wavelength mid-infrared region through commercial distributed feedback quantum cascade lasers. The strongest benzene absorbance band in the infrared is near 14.84 μm, and thus was probed for sensitive benzene detection. Wavelength tuning with multidimensional linear regression were employed to selectively measure benzene, carbon dioxide, and acetylene. Cepstral analysis and wavelength tuning were used to develop a selective sensor for fugitive methane emissions. The sensor was proved to be insensitive to baseline laser intensity imperfections and spectral interference from other present species. In combustion studies, it is desirable to have a diagnostic technique that can detect multiple species simultaneously with high sensitivity, selectivity, and fast time response to validate and improve chemical kinetic mechanisms. A mid-infrared laser sensor was developed for selective and sensitive benzene, toluene, ethylbenzene, and xylenes detection in high-temperature shock tube experiments using deep neural networks. The laser was tuned near 3.3 μm, and an off-axis cavity-enhanced absorption spectroscopy setup was used to enable trace detection. Finally, a novel near-infrared laser-based sensor was developed for water-cut sensing in oil-water flow. The sensor was shown to be immune to the presence of salt and sand in the flow and to temperature variations over 25-60°C. This technique has significant advantages for well and reservoir management, where highly accurate water-cut measurements are required.

This dissertation presents novel methods for addressing important challenges related to the robustness of Deep Neural Networks (DNNs) for 3D understanding and in 3D setups. Our research focuses on two main areas, adversarial robustness on 3D data and setups and the robustness of DNNs to realistic 3D scenarios. One paradigm for 3D understanding is to represent 3D as a set of 3D points and learn functions on this set directly. Our first work, AdvPC, addresses the issue of limited transferability and ease of defense against current 3D point cloud adversarial attacks. By using a point cloud Auto-Encoder to generate more transferable attacks, AdvPC surpasses state-of-the-art attacks by a large margin on 3D point cloud attack transferability. Additionally, AdvPC increases the ability to break defenses by up to 38\% as compared to other baseline attacks on the ModelNet40 dataset. Another paradigm of 3D understanding is to perform 2D processing of multiple images of the 3D data. The second work, MVTN, addresses the problem of selecting viewpoints for 3D shape recognition using a Multi-View Transformation Network (MVTN) to learn optimal viewpoints. It combines MVTN with multi-view approaches leading to state-of-the-art results on standard benchmarks ModelNet40, ShapeNet Core55, and ScanObjectNN. MVTN also improves robustness to realistic scenarios like rotation and occlusion. Our third work analyzes the Semantic Robustness of 2D Deep Neural Networks, addressing the problem of high sensitivity toward semantic primitives in DNNs by visualizing the DNN global behavior as semantic maps and observing the interesting behavior of some DNNs. Additionally, we develop a bottom-up approach to detect robust regions of DNNs for scalable semantic robustness analysis and benchmarking of different DNNs. The fourth work, SADA, showcases the problem of lack of robustness in DNNs specifically for the safety-critical applications of autonomous navigation, beyond the simple classification setup. We present a general framework (BBGAN) for black-box adversarial attacks on trained agents, which covers semantic perturbations to the environment of the agent performing the task. BBGAN is trained to generate failure cases that consistently fool a trained agent on tasks such as object detection, self-driving, and autonomous UAV racing.

Deep Reinforcement Learning (DRL) is a new research area, which integrates deep neural networks into reinforcement learning algorithms. It is revolutionizing the field of AI with high performance in the traditional challenges, such as natural language processing, computer vision etc. The current deep reinforcement learning algorithms enable an end to end learning that utilizes deep neural networks to produce effective actions in complex environments from high dimensional sensory observations, such as raw images. The applications of deep reinforcement learning algorithms are remarkable. For example, the performance of trained agent playing Atari video games is comparable, or even superior to a human player. Current studies mostly focus on training single agent and its interaction with dynamic environments. However, in order to cope with complex real-world scenarios, it is necessary to look into multiple interacting agents and their collaborations on certain tasks. This thesis studies the state-of-the-art deep reinforcement learning algorithms and techniques. Through the experiments conducted in several 2D and 3D game scenarios, we investigate how DRL models can be adapted to train multiple agents cooperating with one another, by communications and physical navigations, and achieving their individual goals on complex tasks. / Djup förstärkningsinlärning (DRL) är en ny forskningsdomän som integrerar djupa neurala nätverk i inlärningsalgoritmer. Det har revolutionerat AI-fältet och skapat höga förväntningar på att lösa de traditionella problemen inom AI-forskningen. I detta examensarbete genomförs en grundlig studie av state-of-the-art inom DRL-algoritmer och DRL-tekniker. Genom experiment med flera 2D- och 3D-spelscenarion så undersöks hur agenter kan samarbeta med varandra och nå sina mål genom kommunikation och fysisk navigering.

Learning behaviour of artificial agents is commonly studied in the framework of Reinforcement Learning. Reinforcement Learning gained increasing popularity in the past years. This is partially due to developments that enabled the possibility to employ complex function approximators, such as deep networks, in combination with the framework. Two of the core challenges in Reinforcement Learning are the correct assignment of credits over long periods of time and dealing with sparse rewards. In this thesis we propose a framework based on the notions of goals to tackle these problems. This work implements several components required to obtain a form of goal-directed behaviour, similar to how it is observed in human reasoning. This includes the representation of a goal space, learning how to set goals and finally how to reach them. The framework itself is build upon the options model, which is a common approach for representing temporally extended actions in Reinforcement Learning. All components of the proposed method can be implemented as deep networks and the complete system can be learned in an end-to-end fashion using standard optimization techniques. We evaluate the approachon a set of continuous control problems of increasing difficulty. We show, that we are able to solve a difficult gathering task, which poses a challenge to state-of-the-art Reinforcement Learning algorithms. The presented approach is furthermore able to scale to complex kinematic agents of the MuJoCo benchmark. / Inlärning av beteende för artificiella agenter studeras vanligen inom Reinforcement Learning.Reinforcement Learning har på senare tid fått ökad uppmärksamhet, detta berordelvis på utvecklingen som gjort det möjligt att använda komplexa funktionsapproximerare, såsom djupa nätverk, i kombination med Reinforcement Learning. Två av kärnutmaningarnainom reinforcement learning är credit assignment-problemet över långaperioder samt hantering av glesa belöningar. I denna uppsats föreslår vi ett ramverk baseratpå delvisa mål för att hantera dessa problem. Detta arbete undersöker de komponentersom krävs för att få en form av målinriktat beteende, som liknar det som observerasi mänskligt resonemang. Detta inkluderar representation av en målrymd, inlärningav målsättning, och till sist inlärning av beteende för att nå målen. Ramverket byggerpå options-modellen, som är ett gemensamt tillvägagångssätt för att representera temporaltutsträckta åtgärder inom Reinforcement Learning. Alla komponenter i den föreslagnametoden kan implementeras med djupa nätverk och det kompletta systemet kan tränasend-to-end med hjälp av vanliga optimeringstekniker. Vi utvärderar tillvägagångssättetpå en rad kontinuerliga kontrollproblem med varierande svårighetsgrad. Vi visar att vikan lösa en utmanande samlingsuppgift, som tidigare state-of-the-art algoritmer har uppvisatsvårigheter för att hitta lösningar. Den presenterade metoden kan vidare skalas upptill komplexa kinematiska agenter i MuJoCo-simuleringar.

Yes / Time-series anomaly detection receives increasing research interest given the growing number of data-rich application domains. Recent additions to anomaly detection methods in research literature include deep neural networks (DNNs: e.g., RNN, CNN, and Autoencoder). The nature and performance of these algorithms in sequence analysis enable them to learn hierarchical discriminative features and time-series temporal nature. However, their performance is affected by usually assuming a Gaussian distribution on the prediction error, which is either ranked, or threshold to label data instances as anomalous or not. An exact parametric distribution is often not directly relevant in many applications though. This will potentially produce faulty decisions from false anomaly predictions due to high variations in data interpretation. The expectations are to produce outputs characterized by a level of confidence. Thus, implementations need the Prediction Interval (PI) that quantify the level of uncertainty associated with the DNN point forecasts, which helps in making better-informed decision and mitigates against false anomaly alerts. An effort has been made in reducing false anomaly alerts through the use of quantile regression for identification of anomalies, but it is limited to the use of quantile interval to identify uncertainties in the data. In this paper, an improve time-series anomaly detection method called deep quantile regression anomaly detection (DQR-AD) is proposed. The proposed method go further to used quantile interval (QI) as anomaly score and compare it with threshold to identify anomalous points in time-series data. The tests run of the proposed method on publicly available anomaly benchmark datasets demonstrate its effective performance over other methods that assumed Gaussian distribution on the prediction or reconstruction cost for detection of anomalies. This shows that our method is potentially less sensitive to data distribution than existing approaches. / Petroleum Technology Development Fund (PTDF) PhD Scholarship, Nigeria (Award Number: PTDF/ ED/PHD/IAT/884/16)

Chromosome classification is one of the essential tasks in karyotyping to diagnose genetic abnormalities like some types of cancers and Down syndrome. Deep convolutional neural networks have been widely used in this task, and the accuracy of classification models is exceptionally critical to such sensitive medical diagnoses. However, it is not always possible to meet the expected accuracy rates for diagnosis. So, it is vital to tell how certain or uncertain a model is with its decision. In our work, we use two metrics, entropy and variance, as uncertainty measurements. Moreover, three additional metrics, fail rate, workload, and tolerance range, are used to measure uncertainty metrics’ quality. Four different non-Bayesian methods: deep ensembles, snapshot ensembles, Test Time Augmentation, and Test Time Dropout, are used in experiments. A negative correlation is observed between the accuracy and the uncertainty estimation; the higher the accuracy of the model, the lower the uncertainty. Densenet121 with deep ensembles as the uncertainty evaluation method and variance as the uncertainty metric gives the best outcomes. Densenet121 provides a wider tolerance range and better separation between uncertain and certain predictions. / Kromosomklassificering är en av de viktigaste uppgifterna i Karyotyping för att diagnostisera genetiska abnormiteter som vissa typer av cancer och Downs syndrom. Deep Convolutional Neural Networks har använts i stor utsträckning i denna uppgift, och noggrannheten hos klassificeringsmodeller är exceptionellt kritisk för sådana känsliga medicinska diagnoser. Det är dock inte alltid möjligt att uppfylla de förväntade noggrannhetsgraderna för diagnos. Så det är viktigt att berätta hur säker eller osäker en modell är med sitt beslut. Forskning har gjorts för att uppskatta osäkerheten med bayesiska metoder och icke-bayesiska neurala nätverk, medan lite är känt om kvaliteten på osäkerhetsuppskattningar. I vårt arbete använder vi två mått, entropi och varians, som osäkerhetsmätningar. Dessutom används ytterligare tre mätvärden, felfrekvens, arbetsbelastning och toleransintervall för att mäta osäkerhetsmätarnas kvalitet. Fyra olika icke-bayesiska metoder: djupa ensembler, ögonblicksbilder, Test Time Augmentation (TTA) och Test Time Dropout (TTD) används i experiment. En negativ korrelation observeras mellan noggrannheten och osäkerhetsuppskattningen; ju högre noggrannhet modellen är, desto lägre är osäkerheten. Densenet121 med djupa ensembler som osäkerhetsutvärderingsmetod och varians som osäkerhetsmätvärdet ger de bästa resultaten. De ger ett bredare toleransintervall och bättre separation mellan osäkra och vissa förutsägelser.

We built a system to participate in the eRisk 2019 T1 Shared Task. The aim of the task was to evaluate systems for early risk prediction on the internet, in particular to identify users suffering from eating disorders as accurately andquickly as possible given their history of Reddit posts in chronological order. In the controlled settings of this task, we also evaluated the performance of three different word representation methods: random indexing, GloVe, and ELMo.We discuss our system’s performance, also in the light of the scores obtained by other teams in the shared task. Our results show that our two-step learning approach was quite successful, and we obtained good scores on the early risk prediction metric ERDE across the board. Contrary to our expectations, we did not observe a clear-cut advantage of contextualized ELMo vectors over the commonly used and much more light-weight GloVevectors. Our best model in terms of F1 score turned out to be a model with GloVe vectors as input to the text classifier and a multi-layer perceptron as user classifier. The best ERDE scores were obtained by the model with ELMo vectors and a multi-layer perceptron. The model with random indexing vectors hit a good balance between precision and recall in the early processing stages but was eventually surpassed by the models with GloVe and ELMo vectors. We put forward some possible explanations for the observed results, as well as proposing some improvements to our system.

La traduction automatique vise à traduire des documents d’une langue à une autre sans l’intervention humaine. Avec l’apparition des réseaux de neurones profonds (DNN), la traduction automatique neuronale(NMT) a commencé à dominer le domaine, atteignant l’état de l’art pour de nombreuses langues. NMT a également ravivé l’intérêt pour la traduction basée sur l’interlangue grâce à la manière dont elle place la tâche dans un cadre encodeur-décodeur en passant par des représentations latentes. Combiné avec la flexibilité architecturale des DNN, ce cadre a aussi ouvert une piste de recherche sur la multimodalité, ayant pour but d’enrichir les représentations latentes avec d’autres modalités telles que la vision ou la parole, par exemple. Cette thèse se concentre sur la traduction automatique multimodale(MMT) en intégrant la vision comme une modalité secondaire afin d’obtenir une meilleure compréhension du langage, ancrée de façon visuelle. J’ai travaillé spécifiquement avec un ensemble de données contenant des images et leurs descriptions traduites, où le contexte visuel peut être utile pour désambiguïser le sens des mots polysémiques, imputer des mots manquants ou déterminer le genre lors de la traduction vers une langue ayant du genre grammatical comme avec l’anglais vers le français. Je propose deux approches principales pour intégrer la modalité visuelle : (i) un mécanisme d’attention multimodal qui apprend à prendre en compte les représentations latentes des phrases sources ainsi que les caractéristiques visuelles convolutives, (ii) une méthode qui utilise des caractéristiques visuelles globales pour amorcer les encodeurs et les décodeurs récurrents. Grâce à une évaluation automatique et humaine réalisée sur plusieurs paires de langues, les approches proposées se sont montrées bénéfiques. Enfin,je montre qu’en supprimant certaines informations linguistiques à travers la dégradation systématique des phrases sources, la véritable force des deux méthodes émerge en imputant avec succès les noms et les couleurs manquants. Elles peuvent même traduire lorsque des morceaux de phrases sources sont entièrement supprimés. / Machine translation aims at automatically translating documents from one language to another without human intervention. With the advent of deep neural networks (DNN), neural approaches to machine translation started to dominate the field, reaching state-ofthe-art performance in many languages. Neural machine translation (NMT) also revived the interest in interlingual machine translation due to how it naturally fits the task into an encoder-decoder framework which produces a translation by decoding a latent source representation. Combined with the architectural flexibility of DNNs, this framework paved the way for further research in multimodality with the objective of augmenting the latent representations with other modalities such as vision or speech, for example. This thesis focuses on a multimodal machine translation (MMT) framework that integrates a secondary visual modality to achieve better and visually grounded language understanding. I specifically worked with a dataset containing images and their translated descriptions, where visual context can be useful forword sense disambiguation, missing word imputation, or gender marking when translating from a language with gender-neutral nouns to one with grammatical gender system as is the case with English to French. I propose two main approaches to integrate the visual modality: (i) a multimodal attention mechanism that learns to take into account both sentence and convolutional visual representations, (ii) a method that uses global visual feature vectors to prime the sentence encoders and the decoders. Through automatic and human evaluation conducted on multiple language pairs, the proposed approaches were demonstrated to be beneficial. Finally, I further show that by systematically removing certain linguistic information from the input sentences, the true strength of both methods emerges as they successfully impute missing nouns, colors and can even translate when parts of the source sentences are completely removed.

We created and analyzed a text classification dataset from freely-available web documents from the United Nation's Sustainable Development Goals. We then used it to train and compare different multi-label text classifiers with the aim of exploring the alternatives for methods that facilitate the search of information of this type of documents. We explored the effectiveness of deep learning and transfer learning in text classification by fine-tuning different pre-trained language representations — Word2Vec, GloVe, ELMo, ULMFiT and BERT. We also compared these approaches against a baseline of more traditional algorithms without using transfer learning. More specifically, we used multinomial Naive Bayes, logistic regression, k-nearest neighbors and Support Vector Machines. We then analyzed the results of our experiments quantitatively and qualitatively. The best results in terms of micro-averaged F1 scores and AUROC are obtained by BERT. However, it is also interesting that the second best classifier in terms of micro-averaged F1 scores is the Support Vector Machines, closely followed by the logistic regression classifier, which both have the advantage of being less computationally expensive than BERT. The results also show a close relation between our dataset size and the effectiveness of the classifiers.

abstract: This Master’s thesis includes the design, integration on-chip, and evaluation of a set of imitation learning (IL)-based scheduling policies: deep neural network (DNN)and decision tree (DT). We first developed IL-based scheduling policies for heterogeneous systems-on-chips (SoCs). Then, we tested these policies using a system-level domain-specific system-on-chip simulation framework [11]. Finally, we transformed them into efficient code using a cloud engine [1] and implemented on a user-space emulation framework [61] on a Unix-based SoC. IL is one area of machine learning (ML) and a useful method to train artificial intelligence (AI) models by imitating the decisions of an expert or Oracle that knows the optimal solution. This thesis's primary focus is to adapt an ML model to work on-chip and optimize the resource allocation for a set of domain-specific wireless and radar systems applications. Evaluation results with four streaming applications from wireless communications and radar domains show how the proposed IL-based scheduler approximates an offline Oracle expert with more than 97% accuracy and 1.20× faster execution time. The models have been implemented as an add-on, making it easy to port to other SoCs. / Dissertation/Thesis / Masters Thesis Computer Engineering 2020