Physical Sciences and Mathematics Commons^™

Extracting Dnn Architectures Via Runtime Profiling On Mobile Gpus, Dong Hyub Kim

Masters Theses

Due to significant investment, research, and development efforts over the past decade, deep neural networks (DNNs) have achieved notable advancements in classification and regression domains. As a result, DNNs are considered valuable intellectual property for artificial intelligence providers. Prior work has demonstrated highly effective model extraction attacks which steal a DNN, dismantling the provider’s business model and paving the way for unethical or malicious activities, such as misuse of personal data, safety risks in critical systems, or spreading misinformation. This thesis explores the feasibility of model extraction attacks on mobile devices using aggregated runtime profiles as a side-channel to leak …

An Efficient Privacy-Preserving Framework For Video Analytics, Tian Zhou

Doctoral Dissertations

With the proliferation of video content from surveillance cameras, social media, and live streaming services, the need for efficient video analytics has grown immensely. In recent years, machine learning based computer vision algorithms have shown great success in various video analytic tasks. Specifically, neural network models have dominated in visual tasks such as image and video classification, object recognition, object detection, and object tracking. However, compared with classic computer vision algorithms, machine learning based methods are usually much more compute-intensive. Powerful servers are required by many state-of-the-art machine learning models. With the development of cloud computing infrastructures, people are able …

Automated Identification And Mapping Of Interesting Mineral Spectra In Crism Images, Arun M. Saranathan

Doctoral Dissertations

The Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) has proven to be an invaluable tool for the mineralogical analysis of the Martian surface. It has been crucial in identifying and mapping the spatial extents of various minerals. Primarily, the identification and mapping of these mineral spectral-shapes have been performed manually. Given the size of the CRISM image dataset, manual analysis of the full dataset would be arduous/infeasible. This dissertation attempts to address this issue by describing an (machine learning based) automated processing pipeline for CRISM data that can be used to identify and map the unique mineral signatures present in …

Data To Science With Ai And Human-In-The-Loop, Gustavo Perez Sarabia

Doctoral Dissertations

AI has the potential to accelerate scientific discovery by enabling scientists to analyze vast datasets more efficiently than traditional methods. For example, this thesis considers the detection of star clusters in high-resolution images of galaxies taken from space telescopes, as well as studying bird migration from RADAR images. In these applications, the goal is to make measurements to answer scientific questions, such as how the star formation rate is affected by mass, or how the phenology of bird migration is influenced by climate change. However, current computer vision systems are far from perfect for conducting these measurements directly. They may …

Policy Gradient Methods: Analysis, Misconceptions, And Improvements, Christopher P. Nota

Doctoral Dissertations

Policy gradient methods are a class of reinforcement learning algorithms that optimize a parametric policy by maximizing an objective function that directly measures the performance of the policy. Despite being used in many high-profile applications of reinforcement learning, the conventional use of policy gradient methods in practice deviates from existing theory. This thesis presents a comprehensive mathematical analysis of policy gradient methods, uncovering misconceptions and suggesting novel solutions to improve their performance. We first demonstrate that the update rule used by most policy gradient methods does not correspond to the gradient of any objective function due to the way the …

Multi-Slam Systems For Fault-Tolerant Simultaneous Localization And Mapping, Samer Nashed

Doctoral Dissertations

Mobile robots need accurate, high fidelity models of their operating environments in order to complete their tasks safely and efficiently. Generating these models is most often done via Simultaneous Localization and Mapping (SLAM), a paradigm where the robot alternatively estimates the most up-to-date model of the environment and its position relative to this model as it acquires new information from its sensors over time. Because robots operate in many different environments with different compute, memory, sensing, and form constraints, the nature and quality of information available to individual instances of different SLAM systems varies substantially. `One-size-fits-all' solutions are thus exceedingly …

Towards Robust Long-Form Text Generation Systems, Kalpesh Krishna

Doctoral Dissertations

Text generation is an important emerging AI technology that has seen significant research advances in recent years. Due to its closeness to how humans communicate, mastering text generation technology can unlock several important applications such as intelligent chat-bots, creative writing assistance, or newer applications like task-agnostic few-shot learning. Most recently, the rapid scaling of large language models (LLMs) has resulted in systems like ChatGPT, capable of generating fluent, coherent and human-like text. However, despite their remarkable capabilities, LLMs still suffer from several limitations, particularly when generating long-form text. In particular, (1) long-form generated text is filled with factual inconsistencies to …

Human-Centered Technologies For Inclusive Collection And Analysis Of Public-Generated Data, Mahmood Jasim

Doctoral Dissertations

The meteoric rise in the popularity of public engagement platforms such as social media, customer review websites, and public input solicitation efforts strives for establishing an inclusive environment for the public to share their thoughts, ideas, opinions, and experiences. Many decisions made at a personal, local, or national scale are often fueled by data generated by the public. As such, inclusive collection, analysis, sensemaking, and utilization of pubic-generated data are crucial to support the exercise of successful decision-making processes. However, people often struggle to engage, participate, and share their opinions due to inaccessibility, the rigidity of traditional public engagement methods, …

Quantifying And Enhancing The Security Of Federated Learning, Virat Vishnu Shejwalkar

Doctoral Dissertations

Federated learning is an emerging distributed learning paradigm that allows multiple users to collaboratively train a joint machine learning model without having to share their private data with any third party. Due to many of its attractive properties, federated learning has received significant attention from academia as well as industry and now powers major applications, e.g., Google's Gboard and Assistant, Apple's Siri, Owkin's health diagnostics, etc. However, federated learning is yet to see widespread adoption due to a number of challenges. One such challenge is its susceptibility to poisoning by malicious users who aim to manipulate the joint machine learning …

Learning To See With Minimal Human Supervision, Zezhou Cheng

Doctoral Dissertations

Deep learning has significantly advanced computer vision in the past decade, paving the way for practical applications such as facial recognition and autonomous driving. However, current techniques depend heavily on human supervision, limiting their broader deployment. This dissertation tackles this problem by introducing algorithms and theories to minimize human supervision in three key areas: data, annotations, and neural network architectures, in the context of various visual understanding tasks such as object detection, image restoration, and 3D generation. First, we present self-supervised learning algorithms to handle in-the-wild images and videos that traditionally require time-consuming manual curation and labeling. We demonstrate that …

Foundations Of Node Representation Learning, Sudhanshu Chanpuriya

Doctoral Dissertations

Low-dimensional node representations, also called node embeddings, are a cornerstone in the modeling and analysis of complex networks. In recent years, advances in deep learning have spurred development of novel neural network-inspired methods for learning node representations which have largely surpassed classical 'spectral' embeddings in performance. Yet little work asks the central questions of this thesis: Why do these novel deep methods outperform their classical predecessors, and what are their limitations? We pursue several paths to answering these questions. To further our understanding of deep embedding methods, we explore their relationship with spectral methods, which are better understood, and show …

Bayesian Structural Causal Inference With Probabilistic Programming, Sam A. Witty

Doctoral Dissertations

Reasoning about causal relationships is central to the human experience. This evokes a natural question in our pursuit of human-like artificial intelligence: how might we imbue intelligent systems with similar causal reasoning capabilities? Better yet, how might we imbue intelligent systems with the ability to learn cause and effect relationships from observation and experimentation? Unfortunately, reasoning about cause and effect requires more than just data: it also requires partial knowledge about data generating mechanisms. Given this need, our task then as computational scientists is to design data structures for representing partial causal knowledge, and algorithms for updating that knowledge in …

Machine Learning Modeling Of Polymer Coating Formulations: Benchmark Of Feature Representation Schemes, Nelson I. Evbarunegbe

Masters Theses

Polymer coatings offer a wide range of benefits across various industries, playing a crucial role in product protection and extension of shelf life. However, formulating them can be a non-trivial task given the multitude of variables and factors involved in the production process, rendering it a complex, high-dimensional problem. To tackle this problem, machine learning (ML) has emerged as a promising tool, showing considerable potential in enhancing various polymer and chemistry-based applications, particularly those dealing with high dimensional complexities.

Our research aims to develop a physics-guided ML approach to facilitate the formulations of polymer coatings. As the first step, this …

Effective And Efficient Transfer Learning In The Era Of Large Language Models, Tu Vu

Doctoral Dissertations

Substantial progress has been made in the field of natural language processing (NLP) due to the advent of large language models (LLMs)—deep neural networks with millions or billions of parameters pre-trained on large amounts of unlabeled data. However, these models have common weaknesses, including degenerate performance in data-scarce scenarios, and substantial computational resource requirements. This thesis aims to develop methods to address these limitations for improved applicability and performance of LLMs in resource-constrained settings with limited data and/or computational resources. To address the need for labeled data in data-scarce scenarios, I present two methods, in Chapter 2 and Chapter 3, …

Graph Representation Learning With Box Embeddings, Dongxu Zhang

Doctoral Dissertations

Graphs are ubiquitous data structures, present in many machine-learning tasks, such as link prediction of products and node classification of scientific papers. As gradient descent drives the training of most modern machine learning architectures, the ability to encode graph-structured data using a differentiable representation is essential to make use of this data. Most approaches encode graph structure in Euclidean space, however, it is non-trivial to model directed edges. The naive solution is to represent each node using a separate "source" and "target" vector, however, this can decouple the representation, making it harder for the model to capture information within longer …

Improving User Experience By Optimizing Cloud Services, Ishita Dasgupta

Doctoral Dissertations

Today, cloud services offer myriads of applications, tailor made for different users in the field of weather, health, finance, entertainment, etc. These services fulfill varying genres of user demands over the Internet. For example, these services can be live (live weather radar, ESPN Live) or on-demand services (weather forecasting, Netflix). While these applications cater to different customer requirements, it is necessary for these services to be efficient with respect to latency, scalability, robustness and quality of experience. These systems need to constantly evolve to provide the best user experience and meet the most current demands of the customer. For instance, …

An Introspective Approach For Competence-Aware Autonomy, Connor Basich

Doctoral Dissertations

Building and deploying autonomous systems in the open world has long been a goal of both the artificial intelligence (AI) and robotics communities. From autonomous driving, to health care, to office assistance, these systems have the potential to transform society and alter our everyday lives. The open world, however, presents numerous challenges that question the typical assumptions made by the models and frameworks often used in contemporary AI and robotics. Systems in the open world are faced with an unconstrained and non-stationary environment with a range of heterogeneous actors that is too complex to be modeled in its entirety. Moreover, …

Data-Driven Modeling And Analytics For Greening The Energy Ecosystem, John Wamburu

Doctoral Dissertations

The energy ecosystem is undergoing a major transition from primarily using carbon-intensive energy sources to greener and renewable sources of energy. For instance, electric vehicles (EVs) are rapidly increasing in popularity thereby eliminating gas-based carbon emissions. Similarly, the increased adoption of solar is injecting greener energy into the grid, thus reducing the grid’s overall carbon footprint. At the same time, the proliferation of networked devices and sensors in the grid is enabling energy usage analysis at fine granularity. In this thesis, I argue that data-driven modeling and analytics applied to energy usage data can facilitate optimal carbon reduction in the …

Rigorous Experimentation For Reinforcement Learning, Scott M. Jordan

Doctoral Dissertations

Scientific fields make advancements by leveraging the knowledge created by others to push the boundary of understanding. The primary tool in many fields for generating knowledge is empirical experimentation. Although common, generating accurate knowledge from empirical experiments is often challenging due to inherent randomness in execution and confounding variables that can obscure the correct interpretation of the results. As such, researchers must hold themselves and others to a high degree of rigor when designing experiments. Unfortunately, most reinforcement learning (RL) experiments lack this rigor, making the knowledge generated from experiments dubious. This dissertation proposes methods to address central issues in …

Learning From Sequential User Data: Models And Sample-Efficient Algorithms, Aritra Ghosh

Doctoral Dissertations

Recent advances in deep learning have made learning representation from ever-growing datasets possible in the domain of vision, natural language processing (NLP), and robotics, among others. However, deep networks are notoriously data-hungry; for example, training language models with attention mechanisms sometimes requires trillions of parameters and tokens. In contrast, we can often access a limited number of samples in many tasks. It is crucial to learn models from these `limited' datasets. Learning with limited datasets can take several forms. In this thesis, we study how to select data samples sequentially such that downstream task performance is maximized. Moreover, we study …

Thermal Transport Across 2d/3d Van Der Waals Interfaces, Cameron Foss

Doctoral Dissertations

Designing improved field-effect-transistors (FETs) that are mass-producible and meet the fabrication standards set by legacy silicon CMOS manufacturing is required for pushing the microelectronics industry into further enhanced technological generations. Historically, the downscaling of feature sizes in FETs has enabled improved performance, reduced power consumption, and increased packing density in microelectronics for several decades. However, many are claiming Moore's law no longer applies as the era of silicon CMOS scaling potentially nears its end with designs approaching fundamental atomic-scale limits -- that is, the few- to sub-nanometer range. Ultrathin two-dimensional (2D) materials present a new paradigm of materials science and …

A Very Small Pond: Discovery Systems That Can Be Used With Folio In Academic Libraries, Jaime Taylor, Aaron Neslin

University Libraries Presentations Series

FOLIO, an open source library services platform, does not have a front end patron interface for searching and using library materials. Any library installing FOLIO will need at least one other software to perform those functions. This article evaluates which systems, in a limited marketplace, are available for academic libraries to use with FOLIO.

Nens, Jaime Taylor

University Libraries Presentations Series

NENS (non-student non-employee) are a group of designations for people who are somehow connected to UMass, but who are neither students nor employees. A person’s NENS designation determines what they have access to at UMass, including at the Libraries.

Labeled Modules In Programs That Evolve, Anil K. Saini

Doctoral Dissertations

Multiple methods have been developed for Inductive Program Synthesis, i.e., synthesizing programs consistent with a set of input-output examples. One such method is genetic programming, which searches for programs with desirable properties from the space of all possible programs through an iterated process of variation and selection that is inspired by natural evolution. Genetic programming has been successful in solving problems from multiple domains. These problems are often challenging because of the range of data types and control structures they require to be solved. Nonetheless, there are many programming problems that are routinely solved by human programmers that cannot be …

Enabling Daily Tracking Of Individual’S Cognitive State With Eyewear, Soha Rostaminia

Doctoral Dissertations

Research studies show that sleep deprivation causes severe fatigue, impairs attention and decision making, and affects our emotional interpretation of events, which makes it a big threat to public safety, and mental and physical well-being. Hence, it would be most desired if we could continuously measure one’s drowsiness and fatigue level, their emotion while making decisions, and assess their sleep quality in order to provide personalized feedback or actionable behavioral suggestions to modulate sleep pattern and alertness levels with the aim of enhancing performance, well-being, and quality of life. While there have been decades of studies on wearable devices, we …

Low Resource Language Understanding In Voice Assistants, Subendhu Rongali

Doctoral Dissertations

Voice assistants such as Amazon Alexa, Apple Siri, and Google Assistant have become ubiquitous. They rely on spoken language understanding, which typically consists of an Automatic Speech Recognition (ASR) component and a Natural Language Understanding (NLU) component. ASR takes user speech as input and generates a text transcription. NLU takes the text transcription as input and generates a semantic parse to identify the requested actions, called intents (play music, turn on lights, etc.) and any relevant entities, called slots (which song to play? which lights to turn on?).

These components require massive amounts of training data to achieve good performance. …

Neural Approaches For Language-Agnostic Search And Recommendation, Hamed Rezanejad Asl Bonab

Doctoral Dissertations

There are significant efforts toward developing better neural approaches for information retrieval problems. However, the vast majority of these studies are conducted using English-only data. In fact, trends and statistics of non-English content and users on the Internet show exponential growth and that novel information retrieval systems need to be language-agnostic; they need to bridge the language barrier between users and content, leverage data from high-resource settings for lower-resourced settings, and be able to extend to new languages and local markets easily. To this end, we focus on search and recommendation as two vital components of information systems. We explore …

Unobtrusive Assessment Of Upper-Limb Motor Impairment Using Wearable Inertial Sensors, Brandon R. Oubre

Doctoral Dissertations

Many neurological diseases cause motor impairments that limit autonomy and reduce health-related quality of life. Upper-limb motor impairments, in particular, significantly hamper the performance of essential activities of daily living, such as eating, bathing, and changing clothing. Assessment of impairment is necessary for tracking disease progression, measuring the efficacy of interventions, and informing clinical decision making. Impairment is currently assessed by trained clinicians using semi-quantitative rating scales that are limited by their reliance on subjective, visual assessments. Furthermore, existing scales are often burdensome to administer and do not capture patients' motor performance in home and community settings, resulting in a …

High-Quality Automatic Program Repair, Manish Motwani

Doctoral Dissertations

Software developers spend significant time and effort fixing bugs. Automatic program repair promises to significantly reduce bug-fixing costs. Program repair requires: fault localization — identifying program elements that cause the bug, patch generation — identifying modifications to those program elements to attempt to repair the bug, and patch validation — verifying that the modification actually repairs the bug. Most automatic program repair techniques use the developer-written tests for the repair process and produce seemingly good patches for 11–19% of the bugs in real-world software. However, most of these patches are not correct, as they overfit to the developer-written tests and …

Answer Similarity Grouping And Diversification In Question Answering Systems, Lakshmi Nair Vikraman

Doctoral Dissertations

The rise in popularity of mobile and voice search has led to a shift in IR from document to passage retrieval for non-factoid questions. Various datasets such as MSMarco, as well as efficient retrieval models have been developed to identify single best answer passages for this task. However, such models do not specifically address questions which could have multiple or alternative answers. In this dissertation, we focus on this new research area that involves studying answer passage relationships and how this could be applied to passage retrieval tasks. We first create a high quality dataset for the answer passage similarity …