research
Is Anyone Paying Attention? Automating Classroom Behavior Classification with Vision-Language Models

Is Anyone Paying Attention? Automating Classroom Behavior Classification with Vision-Language Models

Andrew Franck, Brendan Ng, Ben Fitzgerald, Zane Derrod, Chris Cianci, Chris Craney

Occidental College · Mentor: Chris Cianci(Aug. 2025 - Present)

  • Led research of computer vision pipeline achieving 95% accuracy in automating classroom observation, replacing manual coding with ML-based behavioral analysis of STEM lectures using transformer models
  • Recognizes and classifies over 20 distinct actions according to COPUS evaluation actions
Edge Phoneme Recognition for Children's Speech through Age-Aware Training

Edge Phoneme Recognition for Children's Speech through Age-Aware Training

Joel Walsh, Matthew Arboleda, Ryan Arboleda, Sophie Haak, Sam Hjelmeset, Jose Bustamante Ortiz, Andy Franck, Barry Yang

Occidental College · Mentor: Joel Walsh(Jan. 2026 - Present)

  • Developed age-aware phoneme recognition models optimized for deployment on edge devices
  • Focused on children's speech patterns which differ significantly from adult speech used in standard ASR training
Neural Operators for Excitable Media Dynamics in FitzHugh-Nagumo Systems

Neural Operators for Excitable Media Dynamics in FitzHugh-Nagumo Systems

Andrew Franck, Justin Li

Occidental College · Mentor: Justin Li(Jun. 2025 - Dec. 2025)

  • Developed Fourier Neural Operator model to accelerate FitzHugh-Nagumo reaction-diffusion simulations by several orders of magnitude while maintaining <8% relative L2 error and 0.98 R-squared score
  • Trained on dataset of FHN solutions generated via traditional numerical solver across wide parameter space
Gaussian Process-Based Reinforcement Learning for Autonomous Vehicle Sampling

Gaussian Process-Based Reinforcement Learning for Autonomous Vehicle Sampling

John Lipor, Andrew Franck

Portland State University · Mentor: John Lipor(May. 2025 - Aug. 2025)

  • Assisted in development of reinforcement learning-based policy for adaptive level-set estimation in autonomous underwater vehicles, aiming for efficient seabed classification from ambient acoustic data
  • Modeled spatial similarity using Gaussian processes (GPs), and implemented scalable GP inference using GPyTorch prediction and uncertainty estimation
  • Improved the expert policy using exhaustive lookahead methods and AUC-based terminal cost approximation
Deep Learning for Ocular Biometry from Ultrasound Images

Deep Learning for Ocular Biometry from Ultrasound Images

John Lipor, Hadi Khazaei, Andrew Franck, Danesh Khazaei, John Ng, Faryar Etesami

Oregon Health & Science University · Mentor: John Lipor(May. 2023 - Oct. 2023)

  • Developed a deep learning model to analyze ocular ultrasound images, with shared and independent layers based on GoogLeNet model
  • Our approach demonstrates accurate measurement of axial length, offering an automated alternative to manual biometry measurements in clinical settings
Evaluating Machine Learning Strategies for Geothermal Energy Assessments

Evaluating Machine Learning Strategies for Geothermal Energy Assessments

Portland State University · Mentor: John Lipor(Jan. 2023 - May. 2023)

  • Developed ML model in PyTorch to predict geothermal heat flow residuals based on 28 feature dataset
  • Updated research to analyze topographical maps with vision-based deep learning model
  • Recognized differences in topographical terrain which indicate favorable locations for geothermal power plants