<div>Over the years, the performance demand for graphics applications has been steadily increasing. While upgrading the hardware is one direct solution, the emergence of the new low-level and low-overhead graphics APIs like Vulkan also exposed the possibility of improving rendering performance from the bottom of software implementation.</div><div><br></div><div>Most of the recent years’ middle- to high-end personal computers are equipped with both integrated and discrete GPUs. However, with previous graphics APIs, it is hard to put these two heterogeneous GPUs to work concurrently in the same application without tailored driver support.</div><div><br></div><div>This thesis provides an exploration into the utilization of such heterogeneous GPUs in real-time rendering with the help of Vulkan API. This paper first demonstrates the design and implementation details for the proposed heterogeneous GPUs working model. After that, the paper presents the test of two workload offloading strategies: offloading screen space output workload to the integrated GPU and offloading asynchronous computation workload to the integrated GPU.</div><div><br></div>While this study failed to obtain performance improvement through offloading screen space output workload, it is successful in validating that offloading asynchronous computation workload from the discrete GPU to the integrated GPU can improve the overall system performance. This study proves that it is possible to make use of the integrated and discrete GPUs concurrently in the same application with the help of Vulkan. And offloading asynchronous computation workload from the discrete GPU to the integrated GPU can provide up to 3-4% performance improvement with combinations like UHD Graphics 630 + RTX 2070 Max-Q and HD Graphics 630 + GTX 1050.

Word embeddings are utilized in various natural language processing tasks. Although effective in helping computers learn linguistic patterns employed in natural language, word embeddings also tend to learn unwanted word associations. This affects the performance of NLP tasks, as unwanted word associations propagate and amplify biases. Current word association evaluation methods for word embeddings do not account for changes in word embedding models and training corpora, when creating the rubric for word association evaluation. Current literature also lacks a consistent training and evaluation protocol for comparison of word associations across varying word embedding models and varying training corpora. In order to address this gap in prior literature, this research aims to evaluate different types of word associations, not limited to gender, racial or religious attributes, incorporating and evaluating the diachronic and variable nature of words over text data collected over a period of 200 years. This thesis introduces a framework to track changes in word associations between neutral words (proper nouns) and attributes (adjectives), across different word embedding models, over a temporal dimension, by evaluating clustering tendencies between neutral words (proper nouns) and attributive words (adjectives) over five different word embedding frameworks: Word2vec (CBOW), Word2vec (Skip-gram), GloVe, fastText (CBOW) and fastText (Skip-gram) and 20 decades of text data from 1810s to 2000s. <a>Finally, various cluster level and corpus level measurements will be compared across aforementioned word embedding frameworks, to find how</a> word associations evolve with changes in the embedding model and the training corpus.

Defending models against Natural Language Processing adversarial attacks is a challenge because of the discrete nature of the text dataset. However, given the variety of Natural Language Processing applications, it is important to make text processing models more robust and secure. This paper aims to develop techniques that will help text processing models such as BERT to combat adversarial samples that contain misspellings. These developed models are more robust than off the shelf spelling checkers.

This paper considers problems of finding non‐periodic and periodic schedules in a cycle shop which is a special case of a job shop but an extension of a flow shop. The cycle shop means the machine environment where all jobs have to pass the machines over the same route like in a flow shop but some of the machines in the route can be met more than once. We propose a classification of cycle shops and show that recently studied reentrant flow shops, robotic flow shops, loop reentrant flow shops and V shops are special cases of cycle shops. Problems solvable in polynomial time, pseudopolynomial time, NP‐hard problems and performance guarantee approximations are presented. Related earlier results are surveyed.

<div>There has been a great interest in applying Data Science, Machine Learning, and AI-related technologies in recent years. Industries are adopting these technologies very rapidly, which has enabled them to gather valuable data about their businesses. One such industry that can leverage this data to improve their business's output and quality is the logistics and transport industry. This phenomenon provides an excellent opportunity for companies who rely heavily on air transportation to leverage this data to gain valuable insights and improve their business operations. This thesis is aimed to leverage this data to develop techniques to model complex business processes and design a machine learning-based predictive analytical approach to predict process violations.</div><div>This thesis focused on solving delays in shipment delivery by modeling a prediction technique to predict these delays. The approach presented here was based on real airfreight shipping data, which follows the International Air and Transport Association industry standard for airfreight transportation, to identify shipments at risk of being delayed. By leveraging the shipment process structure, this research presented a new approach that solved the complex event-driven structure of airfreight data that made it difficult to model for predictive analytics.</div><div>By applying different data mining and machine learning techniques, prediction techniques were developed to predict delays in delivering airfreight shipments. The prediction techniques were based on random forest and gradient boosting algorithms. To compare and select the best model, the prediction results were interpreted in the form of six confusion matrix-based performance metrics. The results showed that all the predictors had a high specificity of over 90%, but the sensitivity was low, under 44%. Accuracy was observed to be over 75%, and a geometric mean was between 58% – 64%.</div><div>The performance metrics results provided evidence that our approach could be implemented to develop a prediction technique to model complex business processes. Additionally, an early prediction method was designed to test predictors' performance if complete process information was not available. This proposed method delivered compelling evidence suggesting that early prediction can be achieved without compromising the predictor’s performance.</div>

<p>Over the years, the Internet has become a platform where individuals share their thoughts and personal information. In some cases, these content contain some damaging or sensitive information, which a malicious data collector can leverage to exploit the individual. Nevertheless, what people consider to be sensitive is a relative matter: it not only varies from one person to another but also changes through time. Therefore, it is hard to identify what content is considered sensitive or damaging, from the viewpoint of a malicious entity that does not target specific individuals, rather scavenges the data-sharing platforms to identify sensitive information as a whole. However, the actions that users take to change their privacy preferences or hide their information assists these malicious entities in discovering the sensitive content. </p><p><br></p><p>This thesis offers Crowd Blending techniques to create privacy-preserving systems while maintaining platform utility. In particular, we focus on two privacy tasks for two different data-sharing platforms— i) concealing content deletion on social media platforms and ii) concealing censored information in cryptocurrency blockchains. For the concealment of the content deletion problem, first, we survey the users of social platforms to understand their deletion privacy expectations. Second, based on the users’ needs, we propose two new privacy-preserving deletion mechanisms for the next generation of social platforms. Finally, we compare the effectiveness and usefulness of the proposed mechanisms with the current deployed ones through a user study survey. For the second problem of concealing censored information in cryptocurrencies, we present a provably secure stenography scheme using cryptocurrencies. We show the possibility of hiding censored information among transactions of cryptocurrencies.</p>

Internet goes mobile as billions of users are accessing the Internet through their smartphones. Cellular networks play an essential role in providing “anytime, anywhere” network access as the only large-scale wireless network infrastructure in operation. Mobility support is the salient feature indispensable to ensure seamless Internet connectivity to mobile devices wherever the devices go or are. Cellular network operators deploy a huge number of cell towers over geographical areas each with limited radio coverage. When the device moves out of the radio coverage of its serving cell(s), mobility support is performed to hand over its serving cell(s) to another, thereby ensuring uninterrupted network access.<br>Despite a large success at most places, we uncover that state-of-the-practice mobility support in operational cellular networks suffers from a variety of issues which result in unnecessary performance degradation to mobile devices. In this thesis, we dive into these issues in today’s mobility support and explore possible solutions with no or small changes to the existing network infrastructure.<br>We take a new perspective to study and enhance mobility support. We directly examine, troubleshoot and enhance the underlying procedure of mobility support, instead of higher-layer (application/transport) exploration and optimization in other existing studies. Rather than clean slate network-side solutions, we focus on device-side solutions which are compatible with 3GPP standards and operational network infrastructure, promising immediate benefits without requiring any changes on network side.<br>In particular, we address three technical questions by leveraging the power of the devices. First, how is mobility support performed in reality? We leverage device-side observation to monitor the handoff procedures that happen between the network and the device. We unveil that operator-specific configurations and policies play a decisive role under the standard mechanism and conduct a large-scale measurement study to characterize the extremely complex and diverse handoff configurations used by global operators over the world. Second, what is wrong with the existing mobility support? We conduct model-based reasoning and empirical study to examine network performance issues (e.g., handoff instability and unreachability, missed performance) which are caused by improper handoffs. Finally, how to enhance mobility support? We turn passive devices to proactive devices to enhance mobility support. Specifically, we make a showcase solution which exploits device-side inputs to intervene the default handoff procedure and thus indirectly influence the cell selection decision, thereby improving data speed to mobile devices. All the results in this thesis have been validated or evaluated in reality (over top-tier US carriers like AT&T, Verizon, T-Mobile, some even in global carrier networks).

<div>Federated learning (FL) is a decentralized privacy-preserving learning technique in which clients learn a joint collaborative model through a central aggregator without sharing their data. In this setting, all clients learn a single common predictor (FedAvg), which does not generalize well on each client's local data due to the statistical data heterogeneity among clients. In this paper, we address this problem with PersFL, a discrete two-stage personalized learning algorithm. In the first stage, PersFL finds the optimal teacher model of each client during the FL training phase. In the second stage, PersFL distills the useful knowledge from optimal teachers into each user's local model. The teacher model provides each client with some rich, high-level representation that a client can easily adapt to its local model, which overcomes the statistical heterogeneity present at different clients. We evaluate PersFL on CIFAR-10 and MNIST datasets using three data-splitting strategies to control the diversity between clients' data distributions.</div><div><br></div><div>We empirically show that PersFL outperforms FedAvg and three state-of-the-art personalization methods, pFedMe, Per-FedAvg and FedPer on majority data-splits with minimal communication cost. Further, we study the performance of PersFL on different distillation objectives, how this performance is affected by the equitable notion of fairness among clients, and the number of required communication rounds. We also build an evaluation framework with the following modules: Data Generator, Federated Model Generation, and Evaluation Metrics. We introduce new metrics for the domain of personalized FL, and split these metrics into two perspectives: Performance, and Fairness. We analyze the performance of all the personalized algorithms by applying these metrics to answer the following questions: Which personalization algorithm performs the best in terms of accuracy across all the users?, and Which personalization algorithm is the fairest amongst all of them? Finally, we make the code for this work available at https://tinyurl.com/1hp9ywfa for public use and validation.</div>

<p>With the advance of the technologies of Internet of things, smart devices or virtual personal assistants at home, such as Google Assistant, Apple Siri, and Amazon Alexa, have been widely used to control and access different objects like door lock, blobs, air conditioner, and even bank accounts, which makes our life convenient. Because of its ease for operations, voice control becomes a main interface between users and these smart devices. To make voice control more secure, speaker verification systems have been researched to apply human voice as biometrics to accurately identify a legitimate user and avoid the illegal access. In recent studies, however, it has been shown that speaker verification systems are vulnerable to different security attacks such as replay, voice cloning, and adversarial attacks. Among all attacks, adversarial attacks are the most dangerous and very challenging to defend. Currently, there is no known method that can effectively defend against such an attack in speaker verification systems.</p> <p>The goal of this project is to design and implement a defense system that is simple, light-weight, and effectively against adversarial attacks for speaker verification. To achieve this goal, we study the audio samples from adversarial attacks in both the time domain and the Mel spectrogram, and find that the generated adversarial audio is simply a clean illegal audio with small perturbations that are similar to white noises, but well-designed to fool speaker verification. Our intuition is that if these perturbations can be removed or modified, adversarial attacks can potentially loss the attacking ability. Therefore, we propose to add a plugin-function module to preprocess the input audio before it is fed into the verification system. As a first attempt, we study two opposite plugin functions: denoising that attempts to remove or reduce perturbations and noise-adding that adds small Gaussian noises to an input audio. We show through experiments that both methods can significantly degrade the performance of a state-of-the-art adversarial attack. Specifically, it is shown that denoising and noise-adding can reduce the targeted attack success rate of the attack from 100% to only 56% and 5.2%, respectively. Moreover, noise-adding can slow down the attack 25 times in speed and has a minor effect on the normal operations of a speaker verification system. Therefore, we believe that noise-adding can be applied to any speaker verification system against adversarial attacks. To the best of our knowledge, this is the first attempt in applying the noise-adding method to defend against adversarial attacks in speaker verification systems.</p><br>

This thesis describes a data provenance framework and other associated frameworks for utilizing provenance for data quality and reproducibility. We first identify the requirements for the design of a comprehensive provenance framework which can be applicable to various applications, supports a rich set of provenance metadata, and is interoperable with other provenance management systems. We then design and develop a provenance framework, called SimP, addressing such requirements. Next, we present four prominent applications and investigate how provenance data can be beneficial to such applications. The first application is the quality assessment of access control policies. Towards this, we design and implement the ProFact framework which uses provenance techniques for collecting comprehensive data about actions which were either triggered due to a network context or a user (i.e., a human or a device) action. Provenance data are used to determine whether the policies meet the quality requirements. ProFact includes two approaches for policy analysis: structure-based and classification-based. For the structure-based approach, we design tree structures to organize and assess the policy set efficiently. For the classification-based approach, we employ several classification techniques to learn the characteristics of policies and predict their quality. In addition, ProFact supports policy evolution and the assessment of its impact on the policy quality. The second application is workflow reproducibility. Towards this, we implement ProWS which is a provenance-based architecture for retrieving workflows. Specifically, ProWS transforms data provenance into workflows and then organizes data into a set of indexes to support efficient querying mechanisms. ProWS supports composite queries on three types of search criteria: keywords of workflow tasks, patterns of workflow structure, and metadata about workflows (e.g., how often a workflow was used). The third application is the access control policy reproducibility. Towards this, we propose a novel framework, Polisma, which generates attribute-based access control policies from data, namely from logs of historical access requests and their corresponding decisions. Polisma combines data mining, statistical, and machine learning techniques, and capitalizes on potential context information obtained from external sources (e.g., LDAP directories) to enhance the learning process. The fourth application is the policy reproducibility by utilizing knowledge and experience transferability. Towards this, we propose a novel framework, FLAP, which transfer attribute-based access control policies between different parties in a collaborative environment, while considering the challenges of minimal sharing of data and support policy adaptation to address conflict. All frameworks are evaluated with respect to performance and accuracy.