Engineering Honors Undergraduate Research Projects

Permanent URI for this collection

http://hdl.handle.net/1969.1/175229

Browse

Browsing Engineering Honors Undergraduate Research Projects by Issue Date

Filter results by year or month
Now showing 1 - 8 of 8
  • Thumbnail Image
    Item
    A social computing solution to disaster relief
    (Texas A&M University. Libraries, 2019-05) Yang, Yang; Hu, Xia
    Disaster relief has chronically been a major issue, and various solutions have been presented, attempting to provide the best relief. Currently, disaster rescue teams are facing the problem of lack of valid information at the rescuring scene, resulting in worse relief and more casualties. Data analytics on social network has received success in multiple other application[s] like spam filtering and trend prediction, showing its potential in the field of disaster relief, with a few potential improvements like expanding the size of the dataset and including a more detailed map. The purpose of this research is to expand on previous applications and use social media data to generate a detailed disaster sitution map for first responders. With validated information about the disaster, both survivors and recuers can pinpoint hazardous areas and avoid further damage.
  • Thumbnail Image
    Item
    Performance Study of Graph Convolutional Networks for Medical Prediction-Based Networks
    (Texas A&M University. Libraries, 2019-05) D'Antonio, Benjamin; Jiang, Anxiao
    Predicting the effects of Polypharmacy is a difficult task, and a great amount of money is spent annually remedying the effects of negative drug interactions arising from Polypharmacy. However, Machine Learning can be used to give more accurate predictions than traditional means. In this thesis, we survey current methods of applying Machine Learning to Polypharmacy. We rigorously define a theoretical Polypharmacy problem and design a Graph Convolutional Network that can learn to strongly model our problem. We discuss its performance and offer future steps for generalizing the model to gain a better understanding of the field of Polypharmacy and the potential of Machine Learning to improve it.
  • Thumbnail Image
    Item
    Incorporating haptic features into physics-based simulation
    (Texas A&M University. Libraries, 2019-05) Zhao, Rukai; Sueda, Shinjiro
    In our graphic lab, we have developed many physics-based animations focusing on muscles and we hope to create an interactive interface with tactile feedback so that the users can not only see those physical features but also experience the forces in the muscle line. They will be able to touch on the surface of muscles and feel the muscle texture and they will also be able to drag the muscle line and feel the tension and forces. This is especially important for co-contraction of two opposing muscles, since co-contractions do not produce any motion by changes the stiffness of the joint. Therefore, we used the Geomagic Touch (tm) haptic device for generating the haptic feedbacks and to incorporate OpenHaptics for haptic programming.
  • Thumbnail Image
    Item
    Comparing the performance of different quantum error correcting codes
    (Texas A&M University. Libraries, 2019-05) Pickett, Marshall B.; Klappenecker, Andreas
    This thesis discussess Quantum Error Correction and why it is essential for the development of quantum computation. The introduction will cover the basics of quantum computing and classical error correction. In the second section we will show why Quantum Error Correction is necessary and how different types of error correcting codes help detect and prevent errors. The experiments section will compare the real world performance of diferent types of quantum error correcting codes using a quantum simulator and prevent the results of the experiments. The final section will discuss the conclusions [of] our research.
  • Thumbnail Image
    Item
    EFFICIENT POLYNOMIAL ROOT ISOLATION APPLIED TO COMPUTATIONAL GEOMETRY
    (2019-11-14) Lamkin, Jordan
    Over the last several years, the field of polynomial root isolation has been rapidly improving, but the computational geometry applications have been somewhat unexplored. Here, we have an implementation of a curve intersection engine that showcases the current state-of-the-art in root isolation. The engine is capable of taking two implicitly defined curves and locating their intersection points within some required accuracy. From this work, we can clearly see that root isolation is no longer a significant speed issue in computational geometry. The next issue is really speed of the resultant computation used for variable elimination
  • Thumbnail Image
    Item
    Emergence of Prediction in Delayed Reaching Task Through Neuroevolution
    (2023-12-14) Kang, William; Anand, Chris
    Prediction is an important foundation of cognitive and intelligent behavior. Recent advances in deep learning heavily depend on prediction, in the form of self-supervised learning based on prediction and reinforcement learning (reward prediction). However, how such predictive capabilities emerged from simple organisms has not been investigated fully. Prior works have shown the relationship between input delay and predictive function to compensate for such delay. In this thesis, we investigate the emergence of predictive capabilities in simple evolving neural network controllers, where not only the connection weights but also the network topology evolves. We focus on two main research questions: (1) what fitness criterion promotes predictive behavior? and (2) what changes in the neural network structure correlate with predictive function? To test this, we set up a delayed reaching task, where a two-segment arm is controlled to reach a moving target where the target’s location arrives at the arm’s controller with a period of delay. The arm also has an option of picking up a stick (a tool) to extend its reach. We tested several factors to be included in the fitness function: (1) energy usage, (2) tracking target, and (3) number of tool pick-ups. Our results show that minimizing energy usage is a key to the emergence of 1 prediction. As for the evolved network structure, we found that controllers with more recurrent loops perform better in the task, i.e., tracking the predicted location of the moving target. These results lend us two important insights regarding the evolutionary emergence of prediction: (1) energy minimization is a key driving force, without which random strategy (wasteful in terms of energy) can potentially perform equally, and (2) recurrent loops in the neural network controller not only play the traditional role of memory, but they also serve the purpose of prediction. We expect our results to shed new light on the origin of neural architectures supporting prediction and the energetic constraints.
  • Thumbnail Image
    Item
    Image Scaling Attack Simulation: A Measure of Stealth and Detectability
    (2023-12-14) Kelly, Devon, IV
    Cybersecurity practices require constant effort to be maintained, and one major weakness within the machine learning field is a lack of awareness regarding potential attacks not only in the usage of machine learning models, but in the development process of models as well. It is possible to poison datasets for the benefit of attackers, and for the poor performance of models using data. Previous studies have already determined that preprocessing attacks, such as image scaling attacks, can be difficult to detect both visually and algorithmically. However, there is a lack of emphasis in these studies regarding the real world performance of these attacks and the detectability of the presence of one of these attacks. The purpose of this work is to analyze the relationship between awareness of image scaling attacks with respect to demographic background and experience. We conduct a survey where we gather the subjects’ demographics, analyze the subjects’ experience in cybersecurity, record their responses to a poorly performing convolutional neural network model that has been unknowingly hindered by an image scaling attack of a used dataset, and note their reactions after we reveal to them that the images used within the broken models have been attacked. The subjects in our pilot analysis consist of students taking computer science courses and professors in computer science within Texas A&M University. We find in this study that the overall detection rate of the attack is low enough to be viable in a workplace or academic setting, and that after discovery subjects cannot conclusively determine benign images from attacked images.
  • Thumbnail Image
    Item
    Zenji: Heuristic Algorithms to Identify and Correct Common Japanese Kanji Handwriting Mitakes
    (Texas A&M University. Undergraduate Research. Department of Computer Science and Engineering., 2024-12) Broberg, Cole; Taele, Paul
    The kanji writing system—tens of thousands of logographic Japanese characters originating from China—is by far the most complex of the three Japanese scripts. Kanji present a unique challenge for English-speaking Japanese learners, as they lack an equivalent in western tongues. Whether studying in a classroom or independently, Japanese language learners are often exposed to mnemonics or other techniques to aid in the memorization of these symbols. However, the classroom setting presently provides a unique opportunity for learners to hand-write the characters and receive feedback on their writing. The difficulty and lack of feedback on handwriting for independent learners discourages the practice of hand-writing kanji, which can stunt the learning process. This paper introduces Zenji, a system of heuristic algorithms to analyze student handwriting of kanji. Zenji provides a lightweight solution that may easily be run on students’ personal computing devices and naturally integrated into existing mobile spaced-repetition solutions for kanji learning. The Zenji system provides users overall and per-stroke numeric, visual, and written feedback on various metrics, focusing on common mistakes. In validation testing, Zenji’s feedback proved to be highly accurate across the categories that it assesses, highlighting user errors and combining a classroom-like learning experience with the flexibility of independent study.