There are a plethora of applications that demand high throughput and low latency algorithms leveraging machine learning methods. This need for real time processing can be seen in industries ranging from developing neural network based pre-distortors for enhanced mobile broadband to designing FPGA-based triggers in major scientific efforts by CERN for particle physics. In this thesis, we explore how niche domains can benefit vastly if we look at neurons as a unique boolean function of the form $f:B^{I} \rightarrow B^{O}$, where $B = \{0,1\}$. We focus on how to design topologies that complement such a view of neurons, how to automate such a strategy of neural network design, and inference of such networks on Xilinx FPGAs. Major hardware borne constraints arise when designing topologies that view neurons as unique boolean functions. Fundamentally, realizing such topologies on hardware asserts a strict limit on the 'fan-in' bits of a neuron due to the doubling of permutations possible with every increment in input bit-length. We address this limit by exploring different methods of implementing sparsity and explore activation quantization. Further, we develop a library that supports training a neural network with custom sparsity and quantization. This library also supports conversion of trained Sparse Quantized networks from PyTorch to VERILOG code which is then synthesized using Vivado, all of which is part of the LogicNet tool-flow. To aid faster prototyping, we also support calculation of the worst-case hardware cost of any given topology. We hope that our insights into the behavior of extremely sparse quantized neural networks are of use to the research community and by extension allow people to use the LogicNet design flow to deploy highly efficient neural networks.

Recurrent neural networks (RNNs) are a class of nonlinear dynamical systems often used to model sequence-to-sequence maps. RNNs have been shown to have excellent expressive power but lack stability or robustness guarantees that would be necessary for safety-critical applications. In this paper we formulate convex sets of RNNs with guaranteed stability and robustness properties. The guarantees are derived using differential IQC methods and can ensure contraction (global exponential stability of all solutions) and bounds on incremental l2 gain (the Lipschitz constant of the learnt sequence-to-sequence mapping). An implicit model structure is employed to construct a jointly-convex representation of an RNN and its certificate of stability or robustness. We prove that the proposed model structure includes all previously-proposed convex sets of contracting RNNs as special cases, and also includes all stable linear dynamical systems. We demonstrate the utility of the proposed model class in the context of nonlinear system identification.

Over the last decade, IoT platforms have been developed into a global giant that grabs every aspect of our daily lives by advancing human life with its unaccountable smart services. Because of easy accessibility and fast-growing demand for smart devices and network, IoT is now facing more security challenges than ever before. There are existing security measures that can be applied to protect IoT. However, traditional techniques are not as efficient with the advancement booms as well as different attack types and their severeness. Thus, a strong-dynamically enhanced and up to date security system is required for next-generation IoT system. A huge technological advancement has been noticed in Machine Learning (ML) which has opened many possible research windows to address ongoing and future challenges in IoT. In order to detect attacks and identify abnormal behaviors of smart devices and networks, ML is being utilized as a powerful technology to fulfill this purpose. In this survey paper, the architecture of IoT is discussed, following a comprehensive literature review on ML approaches the importance of security of IoT in terms of different types of possible attacks. Moreover, ML-based potential solutions for IoT security has been presented and future challenges are discussed.

Covariance estimation for matrix-valued data has received an increasing interest in applications including neuroscience and environmental studies. Unlike previous works that rely heavily on matrix normal distribution assumption and the requirement of fixed matrix size, we propose a class of distribution-free regularized covariance estimation methods for high-dimensional matrix data under a separability condition and a bandable covariance structure. Under these conditions, the original covariance matrix is decomposed into a Kronecker product of two bandable small covariance matrices representing the variability over row and column directions. We formulate a unified framework for estimating the banded and tapering covariance, and introduce an efficient algorithm based on rank one unconstrained Kronecker product approximation. The convergence rates of the proposed estimators are studied and compared to the ones for the usual vector-valued data. We further introduce a class of robust covariance estimators and provide theoretical guarantees to deal with the potential heavy-tailed data. We demonstrate the superior finite-sample performance of our methods using simulations and real applications from an electroencephalography study and a gridded temperature anomalies dataset.

"Subject: Please send money Body: I am so distraught. I thought i could reach out to you to help me out. I came down to United Kingdom for a short vacation unfortunately i was mugged at the park of the hotel i stayed, all cash, credit card and cell phone was stolen from me but luckily for me i still have my passport with me. I've been to the embassy and to the police here but they're not helping issues at all and, my flight leaves in few hours time from now but. I am having problems settling the hotel bills and the hotel manager won't let me leave until i settle my hotel bills. I'm freaked out at the moment." As expected, this email, which definitely seems to be spam, ends up in the junk email folder. However, in this paper we show that visual spoofing achieved by substituting some confusables (characters that look similar) into the above email text will enable the same email to bypass the spam filter. We also propose ways to address this loophole.

Factor models are routinely used for dimensionality reduction in modeling of correlated, high-dimensional data. We are particularly motivated by neuroscience applications collecting high-dimensional `predictors' corresponding to brain activity in different regions along with behavioral outcomes. Joint factor models for the predictors and outcomes are natural, but maximum likelihood estimates of these models can struggle in practice when there is model misspecification. We propose an alternative inference strategy based on supervised autoencoders; rather than placing a probability distribution on the latent factors, we define them as an unknown function of the high-dimensional predictors. This mapping function, along with the loadings, can be optimized to explain variance in brain activity while simultaneously being predictive of behavior. In practice, the mapping function can range in complexity from linear to more complex forms, such as splines or neural networks, with the usual tradeoff between bias and variance. This approach yields distinct solutions from a maximum likelihood inference strategy, as we demonstrate by deriving analytic solutions for a linear Gaussian factor model. Using synthetic data, we show that this function-based approach is robust against multiple types of misspecification. We then apply this technique to a neuroscience application resulting in substantial gains in predicting behavioral tasks from electrophysiological measurements in multiple factor models.

While deep neural networks (DNNs) and Gaussian Processes (GPs) are both popularly utilized to solve problems in reinforcement learning, both approaches feature undesirable drawbacks for challenging problems. DNNs learn complex nonlinear embeddings, but do not naturally quantify uncertainty and are often data-inefficient to train. GPs infer posterior distributions over functions, but popular kernels exhibit limited expressivity on complex and high-dimensional data. Fortunately, recently discovered conjugate and neural tangent kernel functions encode the behavior of overparameterized neural networks in the kernel domain. We demonstrate that these kernels can be efficiently applied to regression and reinforcement learning problems by analyzing a baseline case study. We apply GPs with neural network dual kernels to solve reinforcement learning tasks for the first time. We demonstrate, using the well-understood mountain-car problem, that GPs empowered with dual kernels perform at least as well as those using the conventional radial basis function kernel. We conjecture that by inheriting the probabilistic rigor of GPs and the powerful embedding properties of DNNs, GPs using NN dual kernels will empower future reinforcement learning models on difficult domains.

We introduce a novel class of matrices which are defined by the factorization $\textbf{Y} :=\textbf{A}\textbf{X}$, where $\textbf{A}$ is an $m \times n$ wide sparse binary matrix with a fixed number $d$ nonzeros per column and $\textbf{X}$ is an $n \times N$ sparse real matrix whose columns have at most $k$ nonzeros and are $\textit{dissociated}$. Matrices defined by this factorization can be expressed as a sum of $n$ rank one sparse matrices, whose nonzero entries, under the appropriate permutations, form striped blocks - we therefore refer to them as Permuted Striped Block (PSB) matrices. We define the $\textit{PSB data model}$ as a particular distribution over this class of matrices, motivated by its implications for community detection, provable binary dictionary learning with real valued sparse coding, and blind combinatorial compressed sensing. For data matrices drawn from the PSB data model, we provide computationally efficient factorization algorithms which recover the generating factors with high probability from as few as $N =O\left(\frac{n}{k}\log^2(n)\right)$ data vectors, where $k$, $m$ and $n$ scale proportionally. Notably, these algorithms achieve optimal sample complexity up to logarithmic factors.

The proliferation and application of machine learning based Intrusion Detection Systems (IDS) have allowed for more flexibility and efficiency in the automated detection of cyber attacks in Industrial Control Systems (ICS). However, the introduction of such IDSs has also created an additional attack vector; the learning models may also be subject to cyber attacks, otherwise referred to as Adversarial Machine Learning (AML). Such attacks may have severe consequences in ICS systems, as adversaries could potentially bypass the IDS. This could lead to delayed attack detection which may result in infrastructure damages, financial loss, and even loss of life. This paper explores how adversarial learning can be used to target supervised models by generating adversarial samples using the Jacobian-based Saliency Map attack and exploring classification behaviours. The analysis also includes the exploration of how such samples can support the robustness of supervised models using adversarial training. An authentic power system dataset was used to support the experiments presented herein. Overall, the classification performance of two widely used classifiers, Random Forest and J48, decreased by 16 and 20 percentage points when adversarial samples were present. Their performances improved following adversarial training, demonstrating their robustness towards such attacks.

Federated Learning (FL) is a distributed machine learning paradigm where data is decentralized among clients who collaboratively train a model in a computation process coordinated by a central server. By assigning a weight to each client based on the proportion of data instances it possesses, the rate of convergence to an accurate joint model can be greatly accelerated. Some previous works studied FL in a Byzantine setting, where a fraction of the clients may send the server arbitrary or even malicious information regarding their model. However, these works either ignore the issue of data unbalancedness altogether or assume that client weights are known to the server a priori, whereas, in practice, it is likely that weights will be reported to the server by the clients themselves and therefore cannot be relied upon. We address this issue for the first time by proposing a practical weight-truncation-based preprocessing method and demonstrating empirically that it is able to strike a good balance between model quality and Byzantine-robustness. We also establish analytically that our method can be applied to a randomly-selected sample of client weights.