Digital Commons Network^™

Detecting Deepfakes With Deep Learning, Eric C. Tjon

Master's Projects

Advances in generative models and manipulation techniques have given rise to digitally altered videos known as deepfakes. These videos are difficult to identify for both humans and machines. Typical detection methods exploit various imperfections in deepfake videos, such as inconsistent posing and visual artifacts. In this paper, we propose a pipeline with two distinct pathways for examining individual frames and video clips. The image pathway contains a novel architecture called Eff-YNet capable of both segmenting and detecting frames from deepfake videos. It consists of a U-Net with a classification branch and an EfficientNet B4 encoder. The video pathway implements a …

End-To-End Learning Utilizing Temporal Information For Vision- Based Autonomous Driving, Dapeng Guo

Master's Projects

End-to-End learning models trained with conditional imitation learning (CIL) have demonstrated their capabilities in driving autonomously in dynamic environments. The performance of such models however is limited as most of them fail to utilize the temporal information, which resides in a sequence of observations. In this work, we explore the use of temporal information with a recurrent network to improve driving performance. We propose a model that combines a pre-trained, deeper convolutional neural network to better capture image features with a long short-term memory network to better explore temporal information. Experimental results indicate that the proposed model achieves performance gain …

Multi-Agent Deep Reinforcement Learning For Walkers, Inhee Park

Master's Projects

This project was motivated by seeking an AI method towards Artificial General Intelligence (AGI), that is, more similar to learning behavior of human-beings. As of today, Deep Reinforcement Learning (DRL) is the most closer to the AGI compared to other machine learning methods. To better understand the DRL, we compares and contrasts to other related methods: Deep Learning, Dynamic Programming and Game Theory.

We apply one of state-of-art DRL algorithms, called Proximal Policy Op- timization (PPO) to the robot walkers locomotion, as a simple yet challenging environment, inherently continuous and high-dimensional state/action space.

The end goal of this project is …

Variation In Personality Among Semi-Wild Myanmar Timber Elephants, Sateesh Venkatesh

Theses and Dissertations

This study examines two personality traits: exploration and neophobia, which could influence human-elephant conflicts. Thirty-one semi-wild elephants were tested over two trials using a custom novel puzzle tube containing three tasks and three rewards. Our studies show that elephants do vary significantly between individuals in both exploration and neophobia.

A Water Budget And Solute Flux Budget For Waimea River Watershed, Kauai, Hi, U.S.A., Joseph Harold Tolworthy

Theses and Dissertations

Waimea Canyon is a deep V-shaped canyon on the island of Kauai, Hawaii in which the Waimea River and its tributaries flow. The shape and size of the canyon are noteworthy and unusual compared to its contemporary canyons on the Hawaiian Islands which are usually U-shaped or flat bottomed. This could be because there is significantly more physical erosion in Waimea Canyon compared to others. A water budget was created using ArcGIS Pro and data from the University of Hawaii’s rainfall and evapotranspiration atlases, as well as from the United States Geological Survey’s stream gage data. A mass flux was …

Findfur: A Tool For Predicting Furin Cleavage Sites Of Viral Envelope Substrates, Christine Gu

Master's Projects

Most biologically active proteins of eukaryotic cells are initially synthesized in the secretory pathway as inactive precursors and require proteolytic processing to become functionally active. This process is performed by a specialized family of endogenous enzymes known as proproteases convertases (PCs). Within this family of proteases, the most notorious and well-research is furin. Found ubiquitously throughout the human body, typical furin substrates are cleaved at sites composed of paired basic amino acids, specifically at the consensus sequence, R-X-[K/R]-R↓. Furin is often exploited by many pathogens, such as enveloped viruses, for proteolytic processing and maturation of their proteins. Glycoproteins of enveloped …

Optical Signatures Of Plankton In The Open Ocean: From Individual Cells To Global Patterns, Nils Haentjens

Electronic Theses and Dissertations

Marine plankton ecosystems play a major role on Earth, having implications for the global carbon cycle and the food-web structures. Ocean color satellites and networks of autonomous platforms equipped with optical sensors are the primary tools used to study phytoplankton dynamics. They provide long term records while offering a synoptic view of our oceans, enabling to study impact of climate variability on planktonic ecosystems. Interpretation of these observations rely heavily on optical theory and how light propagating through the water is affected by particles who absorb and scatter light (e.g. phytoplankton, sediments). However, the complexity of the optical properties of …

Fostering Coastal Destination Resilience In Maine: Understanding Climate Change Risks And Behaviors, Lydia Horne

Electronic Theses and Dissertations

Tourism is an increasingly important global industry. Coastal and nature-based tourism destinations are especially vulnerable to climate change. Trends in visitation are expected to shift under changing climate conditions, influencing tourist travel behaviors related to destination selection, timing of visits, and activity participation. Tourism suppliers’ adaptation and mitigation behaviors have the potential to alleviate negative shifts in visitation and respond to negative climate change impacts, while also enabling suppliers to take advantage of emerging opportunities. The purpose of this dissertation is to understand how tourism stakeholders, including tourism suppliers (i.e., business owners, managers) and consumers (i.e., visitors), perceive their risk …

Malware Classification With Gaussian Mixture Model-Hidden Markov Models, Jing Zhao

Master's Projects

Discrete hidden Markov models (HMM) are often applied to the malware detection and classification problems. However, the continuous analog of discrete HMMs, that is, Gaussian mixture model-HMMs (GMM-HMM), are rarely considered in the field of cybersecurity. In this study, we apply GMM-HMMs to the malware classification problem and we compare our results to those obtained using discrete HMMs. As features, we consider opcode sequences and entropy-based sequences. For our opcode features, GMM-HMMs produce results that are comparable to those obtained using discrete HMMs, whereas for our entropy-based features, GMM-HMMs generally improve on the classification results that we can attain with …

Bayesian Semi-Supervised Keyphrase Extraction And Jackknife Empirical Likelihood For Assessing Heterogeneity In Meta-Analysis, Guanshen Wang

Statistical Science Theses and Dissertations

This dissertation investigates: (1) A Bayesian Semi-supervised Approach to Keyphrase Extraction with Only Positive and Unlabeled Data, (2) Jackknife Empirical Likelihood Confidence Intervals for Assessing Heterogeneity in Meta-analysis of Rare Binary Events.

In the big data era, people are blessed with a huge amount of information. However, the availability of information may also pose great challenges. One big challenge is how to extract useful yet succinct information in an automated fashion. As one of the first few efforts, keyphrase extraction methods summarize an article by identifying a list of keyphrases. Many existing keyphrase extraction methods focus on the unsupervised setting, …

Analysis Of Github Pull Requests, Canon Ellis

Computer Science and Engineering Theses and Dissertations

The popularity of the software repository site GitHub has created a rise in the Pull Based Development Models' use. An essential portion of pull-based development is the creation of Pull Requests. Pull Requests often have to be reviewed by an individual to be approved and accepted into the Master branch of a software repository. The reviewing process can often be time-consuming and introduce a relatively high level of lost development time. This paper examines thousands of pull requests to understand the most valuable metadata of pull requests. We then introduce metrics in comparing the metadata of pull requests to understand …

Examining Multiple Imputation For Measurement Error Correction In Count Data With Excess Zeros, Shalima Zalsha

Statistical Science Theses and Dissertations

Measurement error and missing data are two common problems in wildlife population surveys. These data are collected from the environment and may be missing or measured with error when the observer’s ability to see the animal is obscured. Methods such as video transects for estimating red snapper abundance and aerial surveys for estimating moose population sizes are highly affected by these problems since total abundance will be underestimated if missing/mismeasured counts are ignored. We shall refer to this problem as visibility bias; it occurs when the true counts are observed when visibility is high, partially observed when visibility is low …

Uncertainty Quantification Of Nonreflecting Boundary Schemes, Brian Citty

Mathematics Theses and Dissertations

Numerical methods have been developed to solve partial differential equations involving the far-field radiation of waves. In addition, there has been recent interest in uncertainty quantification- a burgeoning field involving solving PDEs where random variables are used to model uncertainty in the data. In this thesis we will apply uncertainty quantification methodology to the 1D and 2D wave equation with nonreflecting boundary. We first derive a boundary condition for the 1D wave equation assuming several models of the random wave speed. Later we use our result to compare to an asymptotic SDE approach, and finally we repeat our analysis for …

Analyzing Performance, Energy Consumption, And Reliability Of Mobile Applications, Osama Barack

Computer Science and Engineering Theses and Dissertations

Mobile applications have become a high priority for software developers. Researchers and practitioners are working toward improving and optimizing the energy efficiency and performance of mobile applications due to the capacity limitation of mobile device processors and batteries. In addition, mobile applications have become popular among end-users, developers have introduced a wide range of features that increase the complexity of application code.

To improve and enhance the maintainability, extensibility, and understandability of application code, refactoring techniques were introduced. However, implementing such techniques to mobile applications affects energy efficiency and performance. To evaluate and categorize software implementation and optimization efficiency, several …

Improving A Wireless Localization System Via Machine Learning Techniques And Security Protocols, Zachary Yorio

Masters Theses, 2020-current

The recent advancements made in Internet of Things (IoT) devices have brought forth new opportunities for technologies and systems to be integrated into our everyday life. In this work, we investigate how edge nodes can effectively utilize 802.11 wireless beacon frames being broadcast from pre-existing access points in a building to achieve room-level localization. We explain the needed hardware and software for this system and demonstrate a proof of concept with experimental data analysis. Improvements to localization accuracy are shown via machine learning by implementing the random forest algorithm. Using this algorithm, historical data can train the model and make …

Deep Neural Network Based Student Response Modeling With Uncertainty, Multimodality And Attention, Xinyi Ding

Computer Science and Engineering Theses and Dissertations

In this thesis, I investigate deep neural network based student response modeling, more specifically Knowledge Tracing (KT). Knowledge Tracing allows Intelligent Tutoring Systems to infer which topics or skills a student has mastered, thus adjusting curriculum accordingly. Deep neural network based knowledge tracing models like Deep Knowledge Tracing (DKT) and Dynamic Key-Value Memory Network (DKVMN) have achieved significant improvements compared with conventional probabilistic models. There are mainly two goals in this thesis: 1) To have a better understanding of existing deep neural network based models and their predictions through visualization and through incorporating uncertainties. 2) To improve the performance of …

Multigrid For The Nonlinear Power Flow Equations, Enrique Pereira Batista

Mathematics Theses and Dissertations

The continuously changing structure of power systems and the inclusion of renewable
energy sources are leading to changes in the dynamics of modern power grid,
which have brought renewed attention to the solution of the AC power flow equations.
In particular, development of fast and robust solvers for the power flow problem
continues to be actively investigated. A novel multigrid technique for coarse-graining
dynamic power grid models has been developed recently. This technique uses an
algebraic multigrid (AMG) coarsening strategy applied to the weighted
graph Laplacian that arises from the power network's topology for the construction
of coarse-grain approximations to …

Modeling Fluid Phenomena In The Context Of The Constrained Vapor Bubble System, James Barrett

Mathematics Theses and Dissertations

This thesis focuses on the fluid phenomena observed within what is known as the constrained vapor bubble system. The constrained vapor bubble system is a closed system consisting of a quartz cuvette partially filled with liquid and used as a coolant device. Heat is applied to the heater end which causes the liquid to evaporate and condense on the cooled end of the cuvette. Liquid travels back to the heated end via capillary flow in the corners. A pure vapor bubble is formed in the center of the cuvette giving rise to the name of the experiment. The constrained vapor …

Improved Statistical Methods For Time-Series And Lifetime Data, Xiaojie Zhu

Statistical Science Theses and Dissertations

In this dissertation, improved statistical methods for time-series and lifetime data are developed. First, an improved trend test for time series data is presented. Then, robust parametric estimation methods based on system lifetime data with known system signatures are developed.

In the first part of this dissertation, we consider a test for the monotonic trend in time series data proposed by Brillinger (1989). It has been shown that when there are highly correlated residuals or short record lengths, Brillinger’s test procedure tends to have significance level much higher than the nominal level. This could be related to the discrepancy between …

The Use Of Evidential Reasoning Model With Biomarkers In Pancreatic Cancer Prediction, Qianhui Fan

Master's Projects

In this project, an evidential reasoning model is built to amalgamate factors that could be used in early detection of pancreatic cancer. Our machine learning model outputs a probability of a given patient having prostate cancer based on various input variables. These variables include health history factors, such as smoking and medical history, technical artifacts, such as biopsy sequencing technology, and genomic biomarkers such as mutational, transcriptional and methylomic profiles, cfDNA, and copy number variation. The dataset used in this project is a part of The Cancer Genome Atlas (TCGA) project and was collected from the National Cancer Institute (NIH) …

How Can Employers Contribute To Reducing Commuter-Generated Carbon Emissions? Evaluating Employer-Provided Commuter Benefits In Cambridge, Ma, Mary Richards

Masters Theses

Encouraging a more sustainable commuter mode shift and improving urban transportation systems have the potential to reduce anthropogenic greenhouse gas emissions (GHGs), a major contributor to climate change. Replacing some single-occupancy vehicle (SOV) trips with alternative modes of transportation, such as public transit, walking, or bicycling, represents one approach to begin reducing transportation-related emissions. Collectively, these shifts in transportation patterns would help to reduce the negative social, economic, and environmental costs associated with high rates of personal vehicle use. Employer-provided benefits programs have the potential to influence commuter behavior by making sustainable, alternative commuting choices a more convenient and economically …

Does Invasion Science Encompass The Invaded Range? A Comparison Of The Geographies Of Invasion Science Versus Management In The U.S., Lara Munro

Masters Theses

Biases in invasion science lead to a taxonomic focus on plants, particularly a subset of well-studied plants, and a geographic focus on invasions in Europe and North America. Geographic biases could also cause some branches of invasion science to focus on a subset of environmental conditions in the invaded range, potentially leading to an incomplete understanding of the ecology and management of plant invasions. While broader, country-level geographic biases are well known, it is unclear whether these biases extend to a finer scale and thus affect research within the invaded range. This study assessed whether research sites for ten well-studied …

Metric Learning Via Linear Embeddings For Human Motion Recognition, Byoungdoo Kong

Masters Theses

We consider the application of Few-Shot Learning (FSL) and dimensionality reduction to the problem of human motion recognition (HMR). The structure of human motion has unique characteristics such as its dynamic and high-dimensional nature. Recent research on human motion recognition uses deep neural networks with multiple layers. Most importantly, large datasets will need to be collected to use such networks to analyze human motion. This process is both time-consuming and expensive since a large motion capture database must be collected and labeled. Despite significant progress having been made in human motion recognition, state-of-the-art algorithms still misclassify actions because of characteristics …

Crustal Evolution Of The New England Appalachians: The Rise And Fall Of A Long-Lived Orogenic Plateau, Ian Hillenbrand

Masters Theses

The rise and demise of mountain belts, caused by growth, modification, or removal of the continental lithosphere are fundamental processes that influence almost all Earth systems. Understanding the nature, timing, and significance of active processes in the creation and evolution of modern mountain belts is challenged by a lack of middle crustal and lower crustal exposures. Analogues can be found in ancient orogens, whose deeply eroded roots offer a window into deeper processes, yet this record is complicated by overprinting events and complex deformational histories. Research presented herein constrains the tectonic history of multistage Appalachian Orogen, type locality of the …

Magnetite Mineralization Of The Hammondville Pluton: Poly-Phase Kiruna Type Iocg Magnetite-Apatite Deposits In The Lyon Mountain Granite, Phillip Geer

Masters Theses

Recent mapping of the Eagle Lake Quadrangle, NY, coupled with whole-rock geochemistry and microscopy has offered insight into the petrogenesis of the magnetite-apatite deposits of the Hammondville mining district in the eastern Adirondack Mountains. This study provides insight into the magmatic history of the ca. 1060-1050 Ma Lyon Mountain Granite (Hammondville Pluton) which is intimately related to, and hosts the deposits in this area. Magnetite seams are commonly surrounded by well layered magnetite gneiss, which typically parallel the seams, although in some outcrops appear to be slightly truncated by them. Mineralization is generally concordant with the weak layering found throughout …

Enforcing Higher Standards For Flood Hazard Mitigation In Vermont, Tamsin Flanders

Masters Theses

The state of Vermont faces increasing risk of costly damage from catastrophic flooding events as climate change increases the frequency of heavy rains and cumulative precipitation. In addition to increasing flood inundation risk, extreme precipitation events are leading to high rates damage from fluvial erosion—erosion caused by the force of floodwater and the materials it carries. As in all U.S. states, flood hazard governance in Vermont is shared by multiple levels of government and involves a complex compliance model that relies on local governments to regulate private property owners to achieve community, state, or federal goals.

To encourage municipalities to …

New England’S Underutilized Seafood Species: Defining And Exploring Marketplace Potential In A Changing Climate, Amanda Davis

Masters Theses

New England’s seafood industry has been searching for opportunities to diversify their landings and build resilience as it faces socio-economic challenges from a changing climate. Developing markets for underutilized species is one way the New England community could help their seafood industry build resilience. This thesis identified New England’s underutilized fish species and explored their marketplace potential by examining their availability in a changing climate, current availability to consumers, and consumers’ responses. In Chapter I, I account how New England’s seafood preferences have changed over time. In Chapter II, I identify New England’s seven underutilized seafood species: 1) Acadian redfish …

Designing Surveys On Youth Immigration Reform: Lessons From The 2016 Cces Anomaly, Saige Calkins

Masters Theses

Even with clear advantages to using internet based survey research, there are still some uncertainties to which survey methods are most conducive to an online platform. Most survey method literature, whether focusing on online, telephone, or in-person formats, tend to observe little to no differences between using various survey modes and survey results. Despite this, there is little research focused on the interaction effect between survey formatting, in terms of design and framing, and public opinion on social issues, specifically child immigration policies - a recent topic of popular debate. This paper examines an anomalous result found within the 2016 …

Photothermal And Photochemical Strategies For Lightinduced Shape-Morphing Of Soft Materials, Alexa Simone Kuenstler

Doctoral Dissertations

Engineering materials with the capability to transform energy from photons into mechanical work is an outstanding technical challenge with implications across myriad disciplines. Despite decades of work in this area, comprehensive understanding of how to prescribe shape change and work output in photoactive systems remains limited. To this end, this dissertation explores strategies to assemble photothermal and photochemical moieties in soft material systems to fabricate photoaddressable devices capable of specific shape changes upon illumination. Chapters 2 and 3 describe a methodology for spatially patterning plasmonic nanoparticles in liquid crystal elastomer fibers and sheets to specify local photothermally-induced strain profiles. Using …

Reasoning About User Feedback Under Identity Uncertainty In Knowledge Base Construction, Ariel Kobren

Doctoral Dissertations

Intelligent, automated systems that are intertwined with everyday life---such as Google Search and virtual assistants like Amazon’s Alexa or Apple’s Siri---are often powered in part by knowledge bases (KBs), i.e., structured data repositories of entities, their attributes, and the relationships among them. Despite a wealth of research focused on automated KB construction methods, KBs are inevitably imperfect, with errors stemming from various points in the construction pipeline. Making matters more challenging, new data is created daily and must be integrated with existing KBs so that they remain up-to-date. As the primary consumers of KBs, human users have tremendous potential to …