Computer Sciences Commons^™

Memory Constraints In Cued-Recall-Dependent Learning And Performance Tasks: Why Do Humans Struggle With Simple Yet Memory-Intensive Tasks?, Jack L. Burgess

Dartmouth College Undergraduate Theses

This study explores how various constraints on a computer agent's memory and recall capacities affect how it performs a simple reinforcement learning task: the card-matching memory game "Concentration". Existing computer agents can solve this task easily, but humans struggle with it, even though its rules and objectives are simple. Why is this the case? We identify specific human memory limitations that may be at play: decaying of memories over time and remembering broad characteristics of card locations and faces while forgetting card specifics. Through building and testing a reinforcement learning agent with these human-like memory constraints, we find that they …

Model-Based Deep Siamese Autoencoder For Clustering Single Cell Rna-Seq Data, Zixia Meng

Theses

In the biological field, the smallest unit of organisms in most biological systems is the single cell, and the classification of cells is an everlasting problem. A central task for analysis of single-cell RNA-seq data is to identify and characterize novel cell types. Currently, there are several classical methods, such as K-means algorithm, spectral clustering, and Gaussian Mixture Models (GMMs), which are widely used to cluster the cells. Furthermore, typical dimensional reduction methods such as PCA, t-SNE, and ZIDA have been introduced to overcome “the curse of dimensionality”. A more recent method scDeepCluster has demonstrated improved and promising performances in …

Analysis Of Gameplay Strategies In Hearthstone: A Data Science Approach, Connor W. Watson

Theses

In recent years, games have been a popular test bed for AI research, and the presence of Collectible Card Games (CCGs) in that space is still increasing. One such CCG for both competitive/casual play and AI research is Hearthstone, a two-player adversarial game where players seeks to implement one of several gameplay strategies to defeat their opponent and decrease all of their Health points to zero. Although some open source simulators exist, some of their methodologies for simulated agents create opponents with a relatively low skill level. Using evolutionary algorithms, this thesis seeks to evolve agents with a higher skill …

Ginkgo Biloba Extract Protects Against Methotrexate-Induced Hepatotoxicity: A Computational And Pharmacological Approach, Lina Tariq Al Kury, Fazli Dayyan, Fawad Ali Shah, Zulkifal Malik, Atif Ali Khan Khalil, Abdullah Alattar, Reem Alshaman, Amjad Ali, Zahid Khan

All Works

Ginkgo biloba extract possess several promising biological activities; currently, it is clinically employed in the management of several diseases. This research work aimed to extrapolate the antioxidant and anti-inflammatory effects of Ginkgo biloba (Gb) in methotrexate (MTX)-induced liver toxicity model. These effects were analyzed using different in vivo experimental approaches and by bioinformatics analysis. Male SD rats were grouped as follows: saline; MTX; Gb (pretreated for seven days with 60, 120, and 180 mg/kg daily dose before MTX treatment); silymarin (followed by MTX treatment); Gb 180 mg/kg daily only; and silymarin only. Histopathological results revealed that MTX induced marked hepatic …

Evaluating Driving Performance Of A Novel Behavior Planning Model On Connected Autonomous Vehicles, Keyur Shah

Honors Scholar Theses

Many current algorithms and approaches in autonomous driving attempt to solve the "trajectory generation" or "trajectory following” problems: given a target behavior (e.g. stay in the current lane at the speed limit or change lane), what trajectory should the vehicle follow, and what inputs should the driving agent apply to the throttle and brake to achieve this trajectory? In this work, we instead focus on the “behavior planning” problem—specifically, should an autonomous vehicle change lane or keep lane given the current state of the system?

In addition, current theory mainly focuses on single-vehicle systems, where vehicles do not communicate with …

Regression-Based Motion Planning, Josiah K. Putman

Dartmouth College Undergraduate Theses

This thesis explores two novel approaches to sample-based motion planning that utilize regressions as continuous function approximations to reduce the memory cost of planning. The first approach, Field Search Trees (FST) provides a solution for single-start planning by iteratively building local regressions of the cost-to-arrive function. The second approach, the Regression Complex (RC), constructs a complex of cells with each cell containing a regression of the distance between any two points on its boundary, creating a useful data structure for any start and goal planning query. We provide formal definitions of both approaches and experimental results of running the algorithms …

An Investigation Into The Optimal Use Of Frequency-Based Weights To Improve The Performance Of Entity Resolution, Bingyi Zhong

Theses and Dissertations

Using a weight-based match score has been studied as a way to improve the accuracy of record linking (entity resolution) since Fellegi and Sunter first described the idea of probabilistic agreement and disagreement weights in their seminal work “A Theory of Record Linking.” However, the original work only described weight associated with an entity attribute. Later researchers such as Herzog et al suggested the weighting scheme could be extended to apply to frequently-occurring attribute values (frequency-based weights) instead of just the attribute as in the Fellegi-Sunter scheme. However, there has been little definitive research as to how frequency-based weights should …

Evolution Of Computational Thinking Contextualized In A Teacher-Student Collaborative Learning Environment., John Arthur Underwood

LSU Doctoral Dissertations

The discussion of Computational Thinking as a pedagogical concept is now essential as it has found itself integrated into the core science disciplines with its inclusion in all of the Next Generation Science Standards (NGSS, 2018). The need for a practical and functional definition for teacher practitioners is a driving point for many recent research endeavors. Across the United States school systems are currently seeking new methods for expanding their students’ ability to analytically think and to employee real-world problem-solving strategies (Hopson, Simms, and Knezek, 2001). The need for STEM trained individuals crosses both the vocational certified and college degreed …

Improving Micro-Expression Recognition With Shift Matrices And Database Combination, Yuqi Zhou

Senior Projects - Computer Science & Software Engineering

Micro-expressions are brief, subtle changes in facial expressions associated with emotional responses, and researchers have worked for decades on automatic recognition of them. As convolutional neural networks have been widely used in many areas of computer vision, such as image recognition and motion detection, it has also drawn the attention of scientists to use it for micro-expression recognition. However, none of them have been able to achieve an accuracy high enough for practical use. One of the biggest problems is the limited number of available datasets. The most popular datasets are SMIC, CASME, CASMEII, and SAMM. Most groups have worked …

Reducing Run-Time Adaptation Space Via Analysis Of Possible Utility Bounds, Clay Stevens, Hamid Bagheri

School of Computing: Conference and Workshop Papers

Self-adaptive systems often employ dynamic programming or similar techniques to select optimal adaptations at run-time. These techniques suffer from the “curse of dimensionality", increasing the cost of run-time adaptation decisions. We propose a novel approach that improves upon the state-of-the-art proactive self-adaptation techniques to reduce the number of possible adaptations that need be considered for each run-time adaptation decision. The approach, realized in a tool called Thallium, employs a combination of automated formal modeling techniques to (i) analyze a structural model of the system showing which configurations are reachable from other configurations and (ii) compute the utility that can be …

Machine Learning-Based Signal Degradation Models For Attenuated Underwater Optical Communication Oam Beams, Patrick L. Neary, Abbie T. Watnik, K. Peter Judd, James R. Lindle, Nicholas S. Flann

Computer Science Faculty and Staff Publications

Signal attenuation in underwater communications is a problem that degrades classification performance. Several novel CNN-based (SMART) models are developed to capture the physics of the attenuation process. One model is built and trained using automatic differentiation and another uses the radon cumulative distribution transform. These models are inserted in the classifier training pipeline. It is shown that including these attenuation models in classifier training significantly improves classification performance when the trained model is tested with environmentally attenuated images. The improved classification accuracy will be important in future OAM underwater optical communication applications.

When Agile Means Staying: A Moderated Mediated Model, Tenace Kwaku Setor

Information Systems and Quantitative Analysis Faculty Publications

The design of software development methods focuses on improving task processes, including accommodating changing user requirements and accelerating product delivery. However, there is limited research on how the use of different software development methods impacts IT professionals’ perceptions of organizational mobility. Drawing on concepts from the agile development literature and job characteristics theory, we formulate a moderated mediation model explicating the mechanism and the condition under which agile development use exerts an influence on IT professionals’ intention to stay with their current employer. Specifically, we examine job satisfaction as mediating the effect of using agile development on the intention to …

Ship Detection Feature Analysis In Optical Satellite Imagery Through Machine Learning Applications, Sylvia Charchut

LSU New Orleans Theses and Dissertations

Ship detection remains an important challenge within the government and the commercial industry. Current research has focused on deep learning and has found high success with large labeled datasets. However, deep learning becomes insufficient for limited datasets as well as when explainability is required. There exist scenarios in which explainability and human-in-the-loop processing are needed, such as in naval applications. In these scenarios, handcrafted features and traditional classification algorithms can be useful. This research aims at analyzing multiple textures and statistical features on a small optical satellite imagery dataset. The feature analysis consists of Haar-like features, Haralick features, Hu moments, …

Accelerating The Information-Theoretic Approach Of Community Detection Using Distributed And Hybrid Memory Parallel Schemes, Md Abdul Motaleb Faysal

LSU New Orleans Theses and Dissertations

There are several approaches for discovering communities in a network (graph). Despite being approximating in nature, discovering communities based on the laws of Information Theory has a proven standard of accuracy. The information-theoretic algorithm known as Infomap developed a decade ago for detecting communities, did not foresee the tremendous growth of social networking, multimedia, and massive information boom. To discover communities in massive networks, we have designed a distributed-memory-parallel Infomap in the MPI framework. Our design reaches scalability of over 500 processes capable of processing networks with millions of edges while maintaining quality comparable to the sequential Infomap. We have …

Evidence-Based Detection Of Pancreatic Canc, Rajeshwari Deepak Chandratre

Master's Projects

This study is an effort to develop a tool for early detection of pancreatic cancer using evidential reasoning. An evidential reasoning model predicts the likelihood of an individual developing pancreatic cancer by processing the outputs of a Support Vector Classifier, and other input factors such as smoking history, drinking history, sequencing reads, biopsy location, family and personal health history. Certain features of the genomic data along with the mutated gene sequence of pancreatic cancer patients was obtained from the National Cancer Institute (NIH) Genomic Data Commons (GDC). This data was used to train the SVC. A prediction accuracy of ~85% …

Using Machine Learning To Optimize Predictive Models Used For Big Data Analytics In Various Sports Events, Akhil Kumar Gour

Master's Projects

In today’s world, data is growing in huge volume and type day by day. Historical data can hence be leveraged to predict the likelihood of the events which are to occur in the future. This process of using statistical or any other form of data to predict future outcomes is commonly termed as predictive modelling. Predictive modelling is becoming more and more important and is trending because of several reasons. But mainly, it enables businesses or individual users to gain accurate insights and allows to decide suitable actions for a profitable outcome.

Machine learning techniques are generally used in order …

Predicting Students’ Performance By Learning Analytics, Sandeep Subhash Madnaik

Master's Projects

The field of Learning Analytics (LA) has many applications in today’s technology and online driven education. Learning Analytics is a multidisciplinary topic for learn- ing purposes that uses machine learning, statistic, and visualization techniques [1]. We can harness academic performance data of various components in a course, along with the data background of each student (learner), and other features that might affect his/her academic performance. This collected data then can be fed to a sys- tem with the task to predict the final academic performance of the student, e.g., the final grade. Moreover, it allows students to monitor and self-assess …

Facilitating Mixed Self-Timed Circuits, Alexandra R. Hanson

University Honors Theses

Designers constrain the ordering of computation events in self-timed circuits to ensure the correct behavior of the circuits. Different circuit families utilize different constraints that, when families are combined, may be more difficult to guarantee in combination without inserting delay to postpone necessary events. By analyzing established constraints of different circuit families like Click and GasP, we are able to identify the small changes necessary to either 1) avoid constraints entirely; or 2) decrease the likelihood of necessary delay insertion. Because delay insertion can be tricky for novice designers and because the likelihood of its requirement increases when mixing different …

Inhibition Of Cancer Causing Genes Through The Delivery Of Omomyc In Anti-Myc Therapy: A Systematic Review, Angie Mcgraw

University Honors Theses

A systematic review of the available studies on the interference of OmoMyc with Myc's function in cancerous cells is presented. Myc is a transcription factor that regulates cellular processes such as apoptosis, proliferation, and differentiation. However, Myc is often overexpressed in a variety of cancers, resulting in abnormal growth of cancer cells. Although the inhibition of Myc has been highly desired, it remained a challenge due to its undruggable characteristics. Attempts to inhibit Myc have involved the usage of small-molecules, but these attempts have failed, causing adverse effects and incomplete inhibition of Myc. Despite promising preclinical studies of OmoMyc, it …

Fallen Objects: Collaborating With Artificial Intelligence In The Field Of Graphic Design, Harrison S. Gerard

University Honors Theses

In this paper, I discuss the creation, execution and reception of my digital art series Fallen Objects, in which I collaborate with a neural net to create pseudo-found objects. I explore how artists might collaborate with Artificial Intelligence obliquely, not by having the AI generate the images themselves, but instead generate input for the artists to make the images. While many artists are focused on training neural nets to replicate their own art inputs, I instead focus on working with an AI trained on external, easily-accessible data and creating images from the prompts it delivers. In this way, the AI …

On Session Languages, Prashant Anantharaman, Sean W. Smith

Computer Science Technical Reports

The LangSec approach defends against crafted input attacks by defining a formal language specifying correct inputs and building a parser that decides that language. However, each successive input is not necessarily in the same basic language---e.g., most communication protocols use formats that depend on values previously received, or on some other additional context. When we try to use LangSec in these real-world scenarios, most parsers we write need additional mechanisms to change the recognized language as the execution progresses. This paper discusses approaches researchers have previously taken to build parsers for such protocols and provides formal descriptions of new sets …

Probabilistic And Machine Learning Enhancement To Conn Toolbox, Gayathri Hanuma Ravali Kuppachi

Master's Projects

Clinical depression is a state of mind where the person suffers from persevering and overpowering sorrow. Existing examinations have exhibited that the course of action of arrangement in the brain of patients with clinical depression has a weird framework topology structure. In the earlier decade, resting-state images of the brain have been under the radar a. Specifically, the topological relationship of the brain aligned with graph hypothesis has discovered a strong connection in patients experiencing clinical depression. However, the systems to break down brain networks still have a couple of issues to be unwound. This paper attempts to give a …

Higher-Order Link Prediction Using Graph Embeddings, Neeraj Chavan

Master's Projects

Link prediction is an emerging field that predicts if two nodes in a network are likely to be connected or not in the near future. Networks model real-world systems using pairwise interactions of nodes. However, many of these interactions may involve more than two nodes or entities simultaneously. For example, social interactions often occur in groups of people, research collaborations are among more than two authors, and biological networks describe interactions of a group of proteins. An interaction that consists of more than two entities is called a higher-order structure. Predicting the occurrence of such higher-order structures helps us solve …

Pattern Analysis And Prediction Of Mild Cognitive Impairment Using The Conn Toolbox, Meenakshi Anbukkarasu

Master's Projects

Alzheimer's is an irreversible neurodegenerative disorder described by dynamic psychological and memory defalcation. It has been accounted for that the pervasiveness of Alzheimer's is to increase by 4 times in a few years, where one in every 75 people will have this disorder. Hence, there is a critical requirement for the analysis of Alzheimer's at its beginning stage to diminish the difficulty of the overall medical complications. The initial state of Alzheimer’s is called Mild cognitive impairment (MCI), and hence it is a decent target for premature diagnosis and treatment of Alzheimer's. This project focuses on coordinating numerous imaging modalities …

Detection Of Mild Cognitive Impairment Using Diffusion Compartment Imaging, Matthew Jones

Master's Projects

The result of applying the Neurite Orientation Density and Dispersion Index (NODDI) algorithm to improve the prediction accuracy for patients diagnosed with MCI is reported. Calculations were carried out using a collection of 68 patients (34 control and 34 with MCI) gathered from the Alzheimer’s Disease Neuroimaging Initiative database (ADNI). Patient data includes the use of high-resolution Magnetic Resonance Images as with as Diffusion Tensor Imaging. A Linear Regression accuracy of 83% was observed using the added NODDI summary statistic: Orientation Dispersion Index (ODI). A statistically significant difference in groups was found between control patients and patients with MCI with …

A Robust Structured Tracker Using Local Deep Features, Mohammadreza Javanmardi, Amir Hossein Farzaneh, Xiaojun Qi

Computer Science Faculty and Staff Publications

Deep features extracted from convolutional neural networks have been recently utilized in visual tracking to obtain a generic and semantic representation of target candidates. In this paper, we propose a robust structured tracker using local deep features (STLDF). This tracker exploits the deep features of local patches inside target candidates and sparsely represents them by a set of templates in the particle filter framework. The proposed STLDF utilizes a new optimization model, which employs a group-sparsity regularization term to adopt local and spatial information of the target candidates and attain the spatial layout structure among them. To solve the optimization …

Malware Classification Based On Hidden Markov Model And Word2vec Features, Aparna Sunil Kale

Master's Projects

Malware classification is an important and challenging problem in information security. Modern malware classification techniques rely on machine learning models that can be trained on a wide variety of features, including opcode sequences, API calls, and byte ��-grams, among many others. In this research, we implement hybrid machine learning techniques, where we train hidden Markov models (HMM) and compute Word2Vec encodings based on opcode sequences. The resulting trained HMMs and Word2Vec embedding vectors are then used as features for classification algorithms. Specifically, we consider support vector machine (SVM), ��-nearest neighbor

(��-NN), random forest (RF), and deep neural network (DNN) classifiers. …

Detection And Analysis Of Malware Evolution, Sunhera Barunkumar Paul

Master's Projects

Malware is a malicious software that causes disruption, allows access to unapproved resources, or performs other unauthorized activity. Developing effective malware detection techniques is a critical aspect of information security. One difficulty that arises is that malware often evolves over time, due to changing goals of malware developers, or to counter advances in detection. This evolution can occur through various modifications in malware code. To maintain effective malware detection, it is necessary to detect and analyze malware evolution so that appropriate countermeasures can be taken. We perform a variety of experiments to detect points in time where a malware family …

Word Embedding Techniques For Malware Classification, Aniket Chandak

Master's Projects

Word embeddings are often used in natural language processing as a means to quantify relationships between words. More generally, these same word embedding techniques can be used to quantify relationships between features. In this paper, we conduct a series of experiments that are designed to determine the effectiveness of word embedding in the context of malware classification. First, we conduct experiments where hidden Markov models (HMM) are directly applied to opcode sequences. These results serve to establish a baseline for comparison with our subsequent word embedding experiments. We then experiment with word embedding vectors derived from HMMs— a technique that …

Ai Quantification Of Language Puzzle To Language Learning Generalization, Harita Shroff

Master's Projects

Online language learning applications provide users multiple ways/games to learn a new language. Some of the ways include rearranging words in the foreign language sentences, filling in the blanks, providing flashcards, and many more. Primarily this research focused on quantifying the effectiveness of these games in learning a new language. Secondarily my goal for this project was to measure the effectiveness of exercises for transfer learning in machine translation. Currently, very little research has been done in this field except for the research conducted by the online platforms to provide assurance to their users [12]. Machine learning has been used …