Please forgive the style as this webpage section is currently in the processes of formatting and revision.
David J. Hoxie, PhD
Post-Doctoral Fellow
e-mail:
I aspire to teach, mentor and guide the next generation of students in the application of explainable machine learning in the interdisciplinary nature of tomorrow’s STEM fields. Specifically, my goal is to facilitate the integration of computer/data science with physics, biology, and chemistry, fostering a more interconnected and innovative approach to these disciplines. I aim to equip students with the tools and skills to illuminate the black box of machine learning. I hope to drive my students curiosity by allowing them to explore, compare and contrast various machine learning approaches to data generated by computational methods they may have seen from a junior year course. I also hope to illustrate to students how methods from the physical sciences form the foundational theories behind many classic machine learning papers, and I hope to make these connections clear and accessible to my students such that they are more prepared and equipped with a board set of tools to help them make advancements, work, and discoveries in tomorrows workforce and research labs.
Education
2023 - 2025
Post-Doctoral Fellowship in Engineering
Project I: "Physically Inspired Shallow Graphical Machine Learning Methods s Ground Vehicle Perception Systems"
Co Project II : "Unsupervised Graphical Based Approaches for Optimization and Synchronization of locally distributed Autonomous Ground Vehicle Co-Simulation"
Co Project III: "Minimal Micron Scale Optical Sensing Requirements for Machine Learning Aided Defect Detection in Stress Fracture Surface of Al-xx Based Alloys"
University of Alabama at Birmingham, AL
Co Project II : "Unsupervised Graphical Based Approaches for Optimization and Synchronization of locally distributed Autonomous Ground Vehicle Co-Simulation"
Co Project III: "Minimal Micron Scale Optical Sensing Requirements for Machine Learning Aided Defect Detection in Stress Fracture Surface of Al-xx Based Alloys"
University of Alabama at Birmingham, AL
2014 - 2023
PhD in Physics
Dissertation Title:Machine Learning and Latent Space Representation of Optical Responses in Nanostructures and Thin Films Nanophotonic Structures"
Defense Date: March 20 2023
University of Alabama at Birmingham, AL
Dissertation Title:Machine Learning and Latent Space Representation of Optical Responses in Nanostructures and Thin Films Nanophotonic Structures"
Defense Date: March 20 2023
University of Alabama at Birmingham, AL
2014 - 2019
MS in Physics
University of Alabama at Birmingham, AL
University of Alabama at Birmingham, AL
2009 - 2013
BS in Physics, minors in math and computer science
Cum Laude
Massachusetts College of Liberal Arts, MA
Cum Laude
Massachusetts College of Liberal Arts, MA
Honors
2014-2014
Cum Laude Massachusetts College of Liberal Arts
2013-2013
$\Sigma\Pi\Sigma$ National Physics Honor Society, Massachusetts College of Liberal Arts
2013-2013
Dean's List,Massachusetts College of Liberal Arts
2010-2013
Honors,Massachusetts College of Liberal Arts
2009-2009
Honors,Camden County College
Teaching
Univerisy of Alabama at Birmingham
2023 - 2025
Physics Teaching Fellowship in Engineering Astronomy Labs 101-103,112,118, Undergrad labs consisted of: Hertzsprung Russell Diagrams, orbital occlusion and period measurements, stellar classification,determination of isotropic asteroid belt distributions, spectral classification of gasses, measurement of Hubble Constant via red shift of type II supernova In person Physics I, II Labs, Physics I, II Recitation Determination of spring constants, energy and frictional forces utilizing Pasco system equipment. determination of resistance,voltage,current,electric and magnetic fields, diffraction and diffusion grating spacing "Minimal Micron Scale Optical Sensing Requirements for Machine Learning Aided Defect Detection in Stress Fracture Surface of Al-xx Based Alloys"
Online Physics I Lab,Physics I, II Recitation
Held live lab sections for Q and A for at home IO Device labs. Ran online
recitation for Physics I and II, AL</p>
Pittsfield Public School System
2013 - 2013
Tutored Math and Physics Tutored high school physics and geometry
Massachusetts College of Liberal Arts
2013 - 2013
Tutor lab college level Math and Physics
Tutored physics, trig and calculus
2009 - 2013
Undergrad Teaching Assistant, Java I, II, CS 101
Answered questions and assisted professor in running Java I and II course. Assisted professor with guiding students to navigate office productivity software in CS101.
Camden County College
2008 - 2009
Comp Lab Assistant
Computer lab assistant, to answer technical questions from students, and non specific course questions
Research
2023 - 2025
Post-Doctoral Fellowship in Engineering
Project I:”Physically Inspired Shallow Graphical Machine Learning Methods For Unsupervised Moving Object Tracking and Prediction in Autonomous Ground Vehicle Perception Systems”
Advisor: Dr. Mohammad R Haider
Co-Advisors: Dr. Vladimir Vantsevich, Sam Misko
Explored applications of reservoir computing based graphical neural networks and reinforcement learning for autonomous vehicle systems to aid in the identification, tracking and prediction of moving objects.
Co-Project II:”Unsupervised Graphical Based Approaches for Optimization
and Synchronization of locally distributed Autonomous Ground Vehicle Co-Simulation”
Advisor: Dr. Vladimir Vantsevich
Co-Advisor: Sam Misko
Applied graphical neural network methods to synchronize timing, reduce computational complexity, and maximize accuracy in coupled, distributed simulations of many body systems.
Co-Project III:”Minimal Micron Scale Optical Sensing Requirements for Machine Learning Aided Defect Detection in Optical and Digitized Radiographs of Stress Fracture Surface of Al-xx Based Alloys”
Advisor: Chad Duke P.E Co-Advisor: Sam Misko
Explored relation of camera imagining systems lens combinations to maximize imagining accuracy while retaining low cost performance.
Project I:”Physically Inspired Shallow Graphical Machine Learning Methods For Unsupervised Moving Object Tracking and Prediction in Autonomous Ground Vehicle Perception Systems”
Advisor: Dr. Mohammad R Haider
Co-Advisors: Dr. Vladimir Vantsevich, Sam Misko
Explored applications of reservoir computing based graphical neural networks and reinforcement learning for autonomous vehicle systems to aid in the identification, tracking and prediction of moving objects.
Co-Project II:”Unsupervised Graphical Based Approaches for Optimization
and Synchronization of locally distributed Autonomous Ground Vehicle Co-Simulation”
Advisor: Dr. Vladimir Vantsevich
Co-Advisor: Sam Misko
Applied graphical neural network methods to synchronize timing, reduce computational complexity, and maximize accuracy in coupled, distributed simulations of many body systems.
Co-Project III:”Minimal Micron Scale Optical Sensing Requirements for Machine Learning Aided Defect Detection in Optical and Digitized Radiographs of Stress Fracture Surface of Al-xx Based Alloys”
Advisor: Chad Duke P.E Co-Advisor: Sam Misko
Explored relation of camera imagining systems lens combinations to maximize imagining accuracy while retaining low cost performance.
2023 - 2025
Dissertation Research, University of Alabama at Birmingham Title:”Machine Learning and Latent Space
Representation of Optical Responses in Nanostructures and Thin Films”
Advisor: Dr. Kannatassen Appavoo; Co-Advisor: Dr.Purushotham Bangalore
Spectral signal optimization for nanophotonic spectra by applying unsupervised generative, classification, regression models on UABs Cheaha super computing cluster.
Studied explainable and interoperability of the genetic algorithms and many-body stochastic neighbor and embedding models.
Utilized linear and non-linear entropy based dimensionality and neural reduction algorithms networks such as variational auto encoders.
Optimization of spectra and images via asymmetric training data for shallow linear Mie scatters in varying environments autoencoders.
Representation of Optical Responses in Nanostructures and Thin Films”
Advisor: Dr. Kannatassen Appavoo; Co-Advisor: Dr.Purushotham Bangalore
Spectral signal optimization for nanophotonic spectra by applying unsupervised generative, classification, regression models on UABs Cheaha super computing cluster.
Studied explainable and interoperability of the genetic algorithms and many-body stochastic neighbor and embedding models.
Utilized linear and non-linear entropy based dimensionality and neural reduction algorithms networks such as variational auto encoders.
Optimization of spectra and images via asymmetric training data for shallow linear Mie scatters in varying environments autoencoders.
2018 - 2018
Paid Research Assistanceship, University of Alabama at Birmingham Advisor: Dr. Ryoichi Kawai
Monte-Carlo simulation of diffusion of ARF proteins in Golgi apparatus
2014 - 2014
Research Experience Undergrad, University of Alabama at Birmingham Advisor: Dr. Clayton Simien
Aided set up and install clean room for atomic, molecular and optical physics Setup lab web page, worked on optical laser control software and hardware
Skills
Research Skills
Research Skills
- Effect of test charges in entropic Lorentzian stochastic embedding techniques for classification of nano photonics in various environments
- Distributed computation on Cheaha Super Computing Cluster and Amazon Cloud Services
- Q-Learning using unity MLAgents
- Created a server side JSP Monte Carlo Simulation to simulate radioactive decay using MYSQL
- Research in robotics to use JNI to create a communication path across three different programming languages
- Researched various unsupervised and semi supervised methods to detect and highlight irregularities in human labeled training data
- Researched various unsupervised and semi supervised to repair or be sensitized to erroneous data by ensuring uniformly sparsely sampled data, low number of neurons, and shallow layers to ensure the algorithm learns first order approximations only.
- Developed a method of utilizing tSNE to visually inspect data and better comprehend dynamical phase spaces in a given dataset.
Coded Physics Simulation
- Distributed Cloud Computation calendar based probabilistic event scheduling for simulation radioactive decay model
- Electric Field Wave propagation using GPU shader language
- N-Body Gravitational Simulation
- N-Body Electric and Magnetic Simulation with visualization of fields
- Diffusion-Reaction simulations
Programming Languages and Utilities
- C, C++, C sharp, Fortran, Basic, Cobal
- Python, Numpy, scipy, Jupyter, Matplotlib, Tensorflow, Tourch, Unity MLAgents
- Mathmatica, Matlab, Maple
- HTML, Latex, CSS, mySQL, Javascript, Java, P5JS
- Unreal, Unity, HLSL, Microsoft Direct-X, Microsoft XNA, Open-GL
Machine Learning
- Image and spectra classifiers, with Shallow, Deep, FCN and CNN networks
- Boltzmann Machines, Restricted Boltzmann Machines, Helmholtz Machines
- Explainable approach to machine learning design
- Recurrent Neural networks, Graph Networks
- Reinforcement learning models via Unity MLAgents
- Variational autoencoders,Generative models, Nearest Neighbor clustering
Prior Enrolled Elective Courses
- Graduate College Level Teaching (Part of CRTL series )
- Graduate Grant Writing
- Graduate Math Physics
- Graduate Computational Physics
- Graduate Deep Learning
- Graduate Biological Neural Network Modeling
- Undergrad Intro to Particle Physics
- Graduate Intro to Biophysics
- Graduate Statistical Mechanics
- Graduate Solid State Physics
- Graduate Academic Writing (Part of CRTL series )
- Graduate Academic Communication (Part of CRTL series )
- Undergrad Calc IV and Vector fields
- Undergrad Differential Equations
- Graduate Introduction to Cloud Computing
- Graduate Cloud Computing Security
Physics Visualization for Education
- Wave mechanics using shader techniques in unity, and unreal engine
- polarizability of charged spheres, thermodynamic visualizations for instructive purposes
- Researched how error related to sample size on two physical phenomena, radioactive decay and an infinite square well.
- Use shader based graphical programming to emulate machine learning systems for distributed web platforms for better visualization and understanding of Ising Models, reaction diffusion, decay, and other physical models models
- Simple modeling of photonic imagining systems to demonstrate various optical resolution criterion, and airy discs for visual understanding.
- Simple modeling using noise to create ground truth datasets of fracture surface for visualization and clarification of machine learning resolution and resolvability studies.
- Developed shader based and numpy based statistical photonic modeling to illustrate the temporal dependence of images and sensor detection.
Additional Experience
Presentations
- M R Haider, S Gardner, D J Hoxie, N Bowen, S Misko, J Smereka, P Jayakumar, and V Vantsevich (2024b). “LESN and Auto Encoders”. GVSETS
- D J Hoxie and M R Haider (2024b). “Perception systems in AGV”. GVSETS
- D J Hoxie and V Vantsevich (2024c). “Many body simulations timinig and synchronization for AGV”. ARC Yearly Review
- D J Hoxie, P Bangalore, and K Appavoo (2021). “Using machine learning to optimize optical response of all-dielectric core-shell nanoparticle”. APS March Meeting Abstracts 2021, C60. 002
- D J Hoxie, P Bangalore, and K Appavoo (2021). “Using machine learning to optimize optical response of all-dielectric core-shell nanoparticle”. APS March Meeting Abstracts 2021, C60. 002
- B S Dhami, R P N Tripathi, D J Hoxie, and K Appavoo (2022). “Determining ultrafast carrier dynamics of hybrid perovskites at various stages of nucleation and growth kinetics”. arXiv preprint arXiv:2201.06510
- D J Hoxie, P Bangalore, and K Appavoo (2019). “Using machine learning to optimize optical response of all-dielectric core-shell nanoparticle”. UAB Research Computing Meeting
- D J Hoxie, S Raja, and R Hasan (2019). “Entropy of PRNGs and the accuracy of Monte- Carlo simulations in a publicly distributed computing environment”. iEEE SouthEast Con Huntsville AL
- D J Hoxie (Daniels) and D Cohen (2013). “Accuracy of Monte Carlo Simulations”. Mas- sachusetts College of Liberal Arts Undergraduate Research Conference
- D J Hoxie and Dr. A D Kucher (2012). “A comparative analysis of numerically solving differ- ential equations”. Massachusetts College of Liberal Arts Undergraduate Research Conference
- D J Hoxie and Dr. Wooters (2013). “Measurement of surface waves via diffracting optics”.
- Massachusetts College of Liberal Arts Undergraduate Research Conference
Publications
- D J Hoxie and M R Haider (2024). “Perception systems in AGV”. GVSETS Technical Report
- D J Hoxie and V Vantsevich (2024). “Many body simulations timinig and synchronization for AGV”. GVSETS Technical Report
- D J Hoxie, Steven Gardner, N Bowen, S Misko, M R Haider, J Smereka, P Jayakumar, and V Vantsevich (2024). Planned submission
- S Gardner, D J Hoxie, N Bowen, S Misko, M R Haider, J Smereka, P Jayakumar, and V Vantsevich (2024). “Deep Echostate Autoencoders”. GVSETS
- M R Haider, S Gardner, D J Hoxie, N Bowen, S Misko, J Smereka, P Jayakumar, and V Vantsevich (2024). “LESN and Auto Encoders”. GVSETS
- S Gardner, D J Hoxie, N Bowen, S Misko, M R Haider, J Smereka, P Jayakumar, and V Vantsevich (2024). “Graphical Networks and Object Detection”.GVSETS
- D J Hoxie and V Vantsevich (2024a). Planned submission in final drafting
- D J Hoxie, P Bangalore, and K Appavoo (2023). “Machine Learing of all-dielectric core-shell nanostructures: the critical role of the objective function in inverse design”. RCS Nanoscale
- DJ Hoxie, A Pant, P Bangalore, and K Appavoo (2022) Planned Submission
- D J Hoxie, A Pant, P Bangalore, and K Appavoo (2022) Planned Submission submission to NGP Comp.
- D J Hoxie, A Pant, P Bangalore, and K Appavoo (2022). Planned Submission submission
- D J Hoxie, S Raka, and R Hasan (2019). “Entropy of PRNGs and the accuracy of Monte- Carlo simulations in a publicly distributed computing environment”. iEEE SouthEast Con Huntsville AL
- B Dhami, R Tripathi, D Hoxie, U Tiwari, and K Appavoo (2022). “Understanding nucleation and growth kinetics of hybrid perovskite microstructures using ultrafast spectroscopy.” arXiv preprint arXiv:2201.06510
- BS Dhami, RPN Tripathi, DJ Hoxie, and K Appavoo (2022). “Revealing ultrafast carrier dynamics of hybrid perovskites at various stages of nucleation and growth kinetics”. Advanced Optical Meterials
- BS Dhami, RPN Tripathi, DJ Hoxie, and K Appavoo (2019). “Determining ultrafast carrier dynamics of hybrid perovskites at various stages of nucleation and growth kinetics”. Bulletin of the American Physical Society 66
- D J Hoxie [Daniels] and D Cohen (2013). “Accuracy of Monte Carlo simulations”. Submitted For Review