t-SNE remains one of the most popular embedding techniques for visualizing high-dimensional data. Most standard packages of t-SNE, such as scikit-learn, use the Barnes-Hut t-SNE (BH t-SNE) algorithm for large datasets. However, existing CPU implementations of this algorithm are inefficient. In this work, we accelerate the BH t-SNE on CPUs via cache optimizations, SIMD, parallelizing sequential steps, and improving parallelization of multithreaded steps. Our implementation (Acc-t-SNE) is up to 261x and 4x faster than scikit-learn and the state-of-the-art BH t-SNE implementation from daal4py, respectively, on a 32-core Intel(R) Icelake cloud instance.

A new form of ML knowledge representation with high generalization power is developed and implemented numerically. Initial $\mathit{IN}$ attributes and $\mathit{OUT}$ class label are transformed into the corresponding Hilbert spaces by considering localized wavefunctions. A partially unitary operator optimally converting a state from $\mathit{IN}$ Hilbert space into $\mathit{OUT}$ Hilbert space is then built from an optimization problem of transferring maximal possible probability from $\mathit{IN}$ to $\mathit{OUT}$, this leads to the formulation of a new algebraic problem. Constructed Knowledge Generalizing Operator $\mathcal{U}$ can be considered as a $\mathit{IN}$ to $\mathit{OUT}$ quantum channel; it is a partially unitary rectangular matrix of the dimension $\mathrm{dim}(\mathit{OUT}) \times \mathrm{dim}(\mathit{IN})$ transforming operators as $A^{\mathit{OUT}}=\mathcal{U} A^{\mathit{IN}} \mathcal{U}^{\dagger}$. Whereas only operator $\mathcal{U}$ projections squared are observable $\left\langle\mathit{OUT}|\mathcal{U}|\mathit{IN}\right\rangle^2$ (probabilities), the fundamental equation is formulated for the operator $\mathcal{U}$ itself. This is the reason of high generalizing power of the approach; the situation is the same as for the Schr\"{o}dinger equation: we can only measure $\psi^2$, but the equation is written for $\psi$ itself.

Handwritten character recognition is a hot topic for research nowadays. If we can convert a handwritten piece of paper into a text-searchable document using the Optical Character Recognition (OCR) technique, we can easily understand the content and do not need to read the handwritten document. OCR in the English language is very common, but in the Bengali language, it is very hard to find a good quality OCR application. If we can merge machine learning and deep learning with OCR, it could be a huge contribution to this field. Various researchers have proposed a number of strategies for recognizing Bengali handwritten characters. A lot of ML algorithms and deep neural networks were used in their work, but the explanations of their models are not available. In our work, we have used various machine learning algorithms and CNN to recognize handwritten Bengali digits. We have got acceptable accuracy from some ML models, and CNN has given us great testing accuracy. Grad-CAM was used as an XAI method on our CNN model, which gave us insights into the model and helped us detect the origin of interest for recognizing a digit from an image.

This thesis considers sequential decision problems, where the loss/reward incurred by selecting an action may not be inferred from observed feedback. A major part of this thesis focuses on the unsupervised sequential selection problem, where one can not infer the loss incurred for selecting an action from observed feedback. We also introduce a new setup named Censored Semi Bandits, where the loss incurred for selecting an action can be observed under certain conditions. Finally, we study the channel selection problem in the communication networks, where the reward for an action is only observed when no other player selects that action to play in the round. These problems find applications in many fields like healthcare, crowd-sourcing, security, adaptive resource allocation, among many others. This thesis aims to address the above-described sequential decision problems by exploiting specific structures these problems exhibit. We develop provably optimal algorithms for each of these setups with weak feedback and validate their empirical performance on different problem instances derived from synthetic and real datasets.

Co-clustering is a class of unsupervised data analysis techniques that extract the existing underlying dependency structure between the instances and variables of a data table as homogeneous blocks. Most of those techniques are limited to variables of the same type. In this paper, we propose a mixed data co-clustering method based on a two-step methodology. In the first step, all the variables are binarized according to a number of bins chosen by the analyst, by equal frequency discretization in the numerical case, or keeping the most frequent values in the categorical case. The second step applies a co-clustering to the instances and the binary variables, leading to groups of instances and groups of variable parts. We apply this methodology on several data sets and compare with the results of a Multiple Correspondence Analysis applied to the same data.

Network embedding (NE) approaches have emerged as a predominant technique to represent complex networks and have benefited numerous tasks. However, most NE approaches rely on a homophily assumption to learn embeddings with the guidance of supervisory signals, leaving the unsupervised heterophilous scenario relatively unexplored. This problem becomes especially relevant in fields where a scarcity of labels exists. Here, we formulate the unsupervised NE task as an r-ego network discrimination problem and develop the SELENE framework for learning on networks with homophily and heterophily. Specifically, we design a dual-channel feature embedding pipeline to discriminate r-ego networks using node attributes and structural information separately. We employ heterophily adapted self-supervised learning objective functions to optimise the framework to learn intrinsic node embeddings. We show that SELENE's components improve the quality of node embeddings, facilitating the discrimination of connected heterophilous nodes.

Medical image segmentation methods typically rely on numerous dense annotated images for model training, which are notoriously expensive and time-consuming to collect. To alleviate this burden, weakly supervised techniques have been exploited to train segmentation models with less expensive annotations. In this paper, we propose a novel point-supervised contrastive variance method (PSCV) for medical image semantic segmentation, which only requires one pixel-point from each organ category to be annotated. The proposed method trains the base segmentation network by using a novel contrastive variance (CV) loss to exploit the unlabeled pixels and a partial cross-entropy loss on the labeled pixels. The CV loss function is designed to exploit the statistical spatial distribution properties of organs in medical images and their variance distribution map representations to enforce discriminative predictions over the unlabeled pixels. Experimental results on two standard medical image datasets demonstrate that the proposed method outperforms the state-of-the-art weakly supervised methods on point-supervised medical image semantic segmentation tasks.

Large machine learning models with improved predictions have become widely available in the chemical sciences. Unfortunately, these models do not protect the privacy necessary within commercial settings, prohibiting the use of potentially extremely valuable data by others. Encrypting the prediction process can solve this problem by double-blind model evaluation and prohibits the extraction of training or query data. However, contemporary ML models based on fully homomorphic encryption or federated learning are either too expensive for practical use or have to trade higher speed for weaker security. We have implemented secure and computationally feasible encrypted machine learning models using oblivious transfer enabling and secure predictions of molecular quantum properties across chemical compound space. However, we find that encrypted predictions using kernel ridge regression models are a million times more expensive than without encryption. This demonstrates a dire need for a compact machine learning model architecture, including molecular representation and kernel matrix size, that minimizes model evaluation costs.

Generating new samples from data sets can mitigate extra expensive operations, increased invasive procedures, and mitigate privacy issues. These novel samples that are statistically robust can be used as a temporary and intermediate replacement when privacy is a concern. This method can enable better data sharing practices without problems relating to identification issues or biases that are flaws for an adversarial attack.

Patient triage at emergency departments (EDs) is necessary to prioritize care for patients with critical and time-sensitive conditions. Different tools are used for patient triage and one of the most common ones is the emergency severity index (ESI), which has a scale of five levels, where level 1 is the most urgent and level 5 is the least urgent. This paper proposes a framework for utilizing machine learning to develop an e-triage tool that can be used at EDs. A large retrospective dataset of ED patient visits is obtained from the electronic health record of a healthcare provider in the Midwest of the US for three years. However, the main challenge of using machine learning algorithms is that most of them have many parameters and without optimizing these parameters, developing a high-performance model is not possible. This paper proposes an approach to optimize the hyperparameters of machine learning. The metaheuristic optimization algorithms simulated annealing (SA) and adaptive simulated annealing (ASA) are proposed to optimize the parameters of extreme gradient boosting (XGB) and categorical boosting (CaB). The newly proposed algorithms are SA-XGB, ASA-XGB, SA-CaB, ASA-CaB. Grid search (GS), which is a traditional approach used for machine learning fine-tunning is also used to fine-tune the parameters of XGB and CaB, which are named GS-XGB and GS-CaB. The six algorithms are trained and tested using eight data groups obtained from the feature selection phase. The results show ASA-CaB outperformed all the proposed algorithms with accuracy, precision, recall, and f1 of 83.3%, 83.2%, 83.3%, 83.2%, respectively.