Artificial Intelligence and Robotics Commons^™

Advancing Cyanobacteria Biomass Estimation From Hyperspectral Observations: Demonstrations With Hico And Prisma Imagery, Ryan E. O'Shea, Nima Pahlevan, Brandon Smith, Mariano Bresciani, Todd Egerton, Claudia Giardino, Lin Li, Tim Moore, Antonio Ruiz-Verdu, Steve Ruberg, Stefan G.H. Simis, Richard Stumpf, Diana Vaičiūtė

Biological Sciences Faculty Publications

Retrieval of the phycocyanin concentration (PC), a characteristic pigment of, and proxy for, cyanobacteria biomass, from hyperspectral satellite remote sensing measurements is challenging due to uncertainties in the remote sensing reflectance (∆R_rs) resulting from atmospheric correction and instrument radiometric noise. Although several individual algorithms have been proven to capture local variations in cyanobacteria biomass in specific regions, their performance has not been assessed on hyperspectral images from satellite sensors. Our work leverages a machine-learning model, Mixture Density Networks (MDNs), trained on a large (N = 939) dataset of collocated in situ chlorophyll-a concentrations (Chla), …

Assessment And Learning In Knowledge Spaces (Aleks) Adaptive System Impact On Students' Perception And Self-Regulated Learning Skills, Honda Harati, Laura Sujo-Montes, Chih-Hsiung Tu, Shadow J.W. Armfield, Cherng-Jyh Yen

Educational Leadership & Workforce Development Faculty Publications

Adaptive learning is an educational method that uses computer algorithms and artificial intelligence (AI) to customize learning materials and activities based on each user's model. Adaptive learning has been used for more than 20 years. However, it is still unique, and no other system could bring more or even similar capabilities than the ones adaptive technology offers, including the application of AI, psychology, psychometrics, machine learning, and providing a personalized learning environment. However, there are not many studies on its practicality, usefulness, improving students' learning skills, students' perception, etc., due to the limited number of institutes investing in this new …

Classification Of Chess Games: An Exploration Of Classifiers For Anomaly Detection In Chess, Masudul Hoque

All Graduate Theses, Dissertations, and Other Capstone Projects

Chess is a strategy board game with its inception dating back to the 15th century. The Covid-19 pandemic has led to a chess boom online with 95,853,038 chess games being played during January, 2021 on lichess.com. Along with the chess boom, instances of cheating have also become more rampant. Classifications have been used for anomaly detection in different fields and thus it is a natural idea to develop classifiers to detect cheating in chess. However, there are no specific examples of this, and it is difficult to obtain data where cheating has occurred. So, in this paper, we develop 4 …

Deapsecure Computational Training For Cybersecurity Students: Improvements, Mid-Stage Evaluation, And Lessons Learned, Wirawan Purwanto, Yuming He, Jewel Ossom, Qiao Zhang, Liuwan Zhu, Karina Arcaute, Masha Sosonkina, Hongyi Wu

University Administration Publications

DeapSECURE is a non-degree computational training program that provides a solid high-performance computing (HPC) and big-data foundation for cybersecurity students. DeapSECURE consists of six modules covering a broad spectrum of topics such as HPC platforms, big-data analytics, machine learning, privacy-preserving methods, and parallel programming. In the second year of this program, to improve the learning experience, we implemented a number of changes, such as grouping modules into two broad categories, "big-data" and "HPC"; creating a single cybersecurity storyline across the modules; and introducing post-workshop (optional) "hackshops." Two major goals of these changes are, firstly, to effectively engage students to maintain …

Ship Deck Segmentation In Engineering Document Using Generative Adversarial Networks, Mohammad Shahab Uddin, Raphael Pamie-George, Daron Wilkins, Andres Sousa Poza, Mustafa Canan, Samuel Kovacic, Jiang Li

Engineering Management & Systems Engineering Faculty Publications

Generative adversarial networks (GANs) have become very popular in recent years. GANs have proved to be successful in different computer vision tasks including image-translation, image super-resolution etc. In this paper, we have used GAN models for ship deck segmentation. We have used 2D scanned raster images of ship decks provided by US Navy Military Sealift Command (MSC) to extract necessary information including ship walls, objects etc. Our segmentation results will be helpful to get vector and 3D image of a ship that can be later used for maintenance of the ship. We applied the trained models to engineering documents provided …

Inference Of Surface Velocities From Oblique Time Lapse Photos And Terrestrial Based Lidar At The Helheim Glacier, Franklyn T. Dunbar Ii

Graduate Student Theses, Dissertations, & Professional Papers

Using time dependent observations derived from terrestrial LiDAR and oblique
time-lapse imagery, we demonstrate that a Bayesian approach to glacial motion es-
timation provides a concise way to incorporate multiple data products into a single
motion estimation procedure effectively producing surface velocity estimates with
an associated uncertainty. This approach brings both improved computational effi-
ciency, and greater scalability across observational time-frames when compared to
existing methods. To gauge efficacy, we apply these methods to a set of observa-
tions from the Helheim Glacier, a critical actor in contemporary mass loss trends
observed in the Greenland Ice Sheet. We find that …

Clustering Data To Classify Hearthstone Decks, Tim Inzitari

Williams Honors College, Honors Research Projects

The esports game of "Hearthstone" is a collectible card game with a competitive format that has every team submit 4 decks of 30 cards each. Using K-Means clustering an adaptable way to group data for classifying can be made that works well in every update of the game. This system will take in a list of decks and cluster them to easily classify large amounts of information in a timely fashion. This system will be able to be used by the Universities esports department for years to come to aid the preparation of "Hearthstone" matches. This model uses qualities about …

Source Code Comment Classification Artificial Intelligence, Cole Sutyak

Williams Honors College, Honors Research Projects

Source code comment classification is an important problem for future machine learning solutions. In particular, supervised machine learning solutions that have largely subjective data labels but are difficult to obtain the labels for. Machine learning problems are problems largely because of a lack of data. In machine learning solutions, it is better to have a large amount of mediocre data than it is to have a small amount of good data. While the mediocre data might not produce the best accuracy, it produces the best results because there is much more to learn from the problem.

In this project, data …

Learning From Multi-Class Imbalanced Big Data With Apache Spark, William C. Sleeman Iv

Theses and Dissertations

With data becoming a new form of currency, its analysis has become a top priority in both academia and industry, furthering advancements in high-performance computing and machine learning. However, these large, real-world datasets come with additional complications such as noise and class overlap. Problems are magnified when with multi-class data is presented, especially since many of the popular algorithms were originally designed for binary data. Another challenge arises when the number of examples are not evenly distributed across all classes in a dataset. This often causes classifiers to favor the majority class over the minority classes, leading to undesirable results …

Detecting Surface Interactions Via A Wearable Microphone To Improve Augmented Reality Text Entry, R. Habibi

Dissertations, Master's Theses and Master's Reports

This thesis investigates whether we can detect and distinguish between surface interaction events such as tapping or swiping using a wearable mic from a surface. Also, what are the advantages of new text entry methods such as tapping with two fingers simultaneously to enter capital letters and punctuation? For this purpose, we conducted a remote study to collect audio and video of three different ways people might interact with a surface. We also built a CNN classifier to detect taps. Our results show that we can detect and distinguish between surface interaction events such as tap or swipe via a …

Weakly Supervised Learning For Multi-Image Synthesis, Muhammad Usman Rafique

Theses and Dissertations--Electrical and Computer Engineering

Machine learning-based approaches have been achieving state-of-the-art results on many computer vision tasks. While deep learning and convolutional networks have been incredibly popular, these approaches come at the expense of huge amounts of labeled data required for training. Manually annotating large amounts of data, often millions of images in a single dataset, is costly and time consuming. To deal with the problem of data annotation, the research community has been exploring approaches that require less amount of labelled data.

The central problem that we consider in this research is image synthesis without any manual labeling. Image synthesis is a classic …

Scaling Up Exact Neural Network Compression By Relu Stability, Thiago Serra, Xin Yu, Abhinav Kumar, Srikumar Ramalingam

Faculty Conference Papers and Presentations

We can compress a rectifier network while exactly preserving its underlying functionality with respect to a given input domain if some of its neurons are stable. However, current approaches to determine the stability of neurons with Rectified Linear Unit (ReLU) activations require solving or finding a good approximation to multiple discrete optimization problems. In this work, we introduce an algorithm based on solving a single optimization problem to identify all stable neurons. Our approach is on median 183 times faster than the state-of-art method on CIFAR-10, which allows us to explore exact compression on deeper (5 x 100) and wider …

The Enlightening Role Of Explainable Artificial Intelligence In Chronic Wound Classification, Salih Sarp, Murat Kuzlu, Emmanuel Wilson, Umit Cali, Ozgur Guler

Engineering Technology Faculty Publications

Artificial Intelligence (AI) has been among the most emerging research and industrial application fields, especially in the healthcare domain, but operated as a black-box model with a limited understanding of its inner working over the past decades. AI algorithms are, in large part, built on weights calculated as a result of large matrix multiplications. It is typically hard to interpret and debug the computationally intensive processes. Explainable Artificial Intelligence (XAI) aims to solve black-box and hard-to-debug approaches through the use of various techniques and tools. In this study, XAI techniques are applied to chronic wound classification. The proposed model classifies …

Food/Non-Food Classification Of Real-Life Egocentric Images In Low- And Middle-Income Countries Based On Image Tagging Features, Guangzong Chen, Wenyan Jia, Yifan Zhao, Zhi-Hong Mao, Benny Lo, Alex K Anderson, Gary Frost, Modou L Jobarteh, Megan A Mccrory, Edward Sazonov, Matilda Steiner-Asiedu, Richard S Ansong, Thomas Baranowski, Lora Burke, Mingui Sun

Children’s Nutrition Research Center Staff Publications

Malnutrition, including both undernutrition and obesity, is a significant problem in low- and middle-income countries (LMICs). In order to study malnutrition and develop effective intervention strategies, it is crucial to evaluate nutritional status in LMICs at the individual, household, and community levels. In a multinational research project supported by the Bill & Melinda Gates Foundation, we have been using a wearable technology to conduct objective dietary assessment in sub-Saharan Africa. Our assessment includes multiple diet-related activities in urban and rural families, including food sources (e.g., shopping, harvesting, and gathering), preservation/storage, preparation, cooking, and consumption (e.g., portion size and nutrition analysis). …

Lstm-Based Model For Human Brain Decisions Using Eeg Signals Analysis, Lorela Bano

College of Graduate Studies: Theses & Dissertations

As machine learning models become more sophisticated, and biometric data becomes more readily available through new non-invasive technologies, it becomes increasingly possible to gain access to interesting biometric data that could revolutionize Human Computer Interaction. In this research, we propose a framework to assess and quantify human preference (like or dislike) on presenting various external visual stimuli. Our framework relies on an Long Short-Term Memory (LSTM) Recurrent Neural Network (RNN) based model and on electroencephalogram (EEG) signals analysis to predict Like or Dislike preference of human subjects when presented with various marketing images.

Implication Of Manifold Assumption In Deep Learning Models For Computer Vision Applications, Marzieh Edraki

Electronic Theses and Dissertations, 2020-2023

The Deep Neural Networks (DNN) have become the main contributor in the field of machine learning (ML). Specifically in the computer vision (CV), there are applications like image and video classification, object detection and tracking, instance segmentation and visual question answering, image and video generation are some of the applications from many that DNNs have demonstrated magnificent progress. To achieve the best performance, the DNNs usually require a large number of labeled samples, and finding the optimal solution for such complex models with millions of parameters is a challenging task. It is known that, the data are not uniformly distributed …

Visual Learning Beyond Human Curated Datasets, Muhammad Abdullah Jamal

Electronic Theses and Dissertations, 2020-2023

The success of deep neural networks in a variety of computer vision tasks heavily relies on large- scale datasets. However, it is expensive to manually acquire labels for large datasets. Given the human annotation cost and scarcity of data, the challenge is to learn efficiently with insufficiently labeled data. In this dissertation, we propose several approaches towards data-efficient learning in the context of few-shot learning, long-tailed visual recognition, and unsupervised and semi-supervised learning. In the first part, we propose a novel paradigm of Task-Agnostic Meta- Learning (TAML) algorithms to improve few-shot learning. Furthermore, in the second part, we analyze the …

Ensemble Encoder-Decoder Models For Predicting Land Transformation, Pariya Pourmohammadi

Graduate Theses, Dissertations, and Problem Reports (ETD)

In studying dynamic and complex processes which are influenced by a system of inter-connected driving variables, it is crucial to apply models that can learn the complexity of the interactions. Land transformation is one of such complex processes, prediction of which can help to mitigate severe climate situations and improve the resiliency of communities. In this study, a multi-spectral set of data cubes is used to capture various characteristics of a geographic region. Based on the data cube, a feature space is constructed using socio-economic attributes, terrain characteristics, and landscape traits of the study region. Two-dimensional and three-dimensional convolutional neural …

Deep Fingerprint Matching From Contactless To Contact Fingerprints For Increased Interoperability, Alexander James Wilson

Graduate Theses, Dissertations, and Problem Reports (ETD)

Contactless fingerprint matching is a common form of biometric security today. Most smartphones and associated apps now let users opt into using this form of biometric security. However, it’s difficult to match a finger-photo to a fingerprint because of perspective distortion occurring at the edges of the finger-photo, so direct matching using conventional methods will not be as accurate due to a lack of sufficient matching minutiae points. To address this issue, we propose a deep model, Perspective Distortion Rectification Model (PDRM), to estimate the fingerprint correspondence for finger-photo images in order to recover more minutiae points. Not only do …

Maidrl: Semi-Centralized Multi-Agent Reinforcement Learning Using Agent Influence, Anthony Lee Harris

Graduate Theses/Dissertations

In recent years, reinforcement learning algorithms, a subset of machine learning that focuses on solving problems through trial-and-error learning, have been used in the field of multi-agent systems to help the agents with interactions and cooperation on a variety of tasks. Given the enormous success of reinforcement learning in single-agent systems like Chess, Shogi, and Go, it is natural for the next step to be the expansion into multi-agent systems. However, controlling multiple agents simultaneously is extremely challenging, as the complexity increases tremendously with the number of agents in the system. Existing approaches in this regard use a wide range …