<Who is Jin Park?>
Jin Hyun Park is a Ph.D. student in Computer Science and Engineering at Texas A&M University, where he works as both a graduate and teaching assistant in the Brain Networks Lab. His research spans computational neuroscience, machine learning, and deep learning, with particular emphasis on how convolutional neural networks mirror human visual perception and on the "Evolution of Prediction," which studies the factors that give rise to 'memory' thus leading to 'prediction'.
He earned his bachelor's degree in Computer Science from the University of Bristol, graduating with First Class Honors. His bachelor thesis "Representations learnt by SGD and Adaptive learning rules: Conditions that vary sparsity and selectivity in neural network" investigated specific circumstances that naturally increase sparsity and selectivity in a simple neural network.
Before joining Texas A&M, he worked at Hyundai Motors Company - Commercial Vehicle Business Innovation Team (Seoul, South Korea) and served as a research assistant at Kyungpook National University - Multidomal Language and Cognition Lab (Daegu, South Korea). After joining Texas A&M University, his research has focused on (1) developing biologically inspired convolutional neural networks that mimic human visual perception and (2) evolution of prediction. During his internship at Tilda Corp. as a machine learning engineer in the Summer of 2023, he optimized neural networks using genetic algorithms. He also interned as a machine learning engineer at LG Uplus (formerly LG Telecom) and Samsung Life Insurance in the Summer of 2024. At LG Uplus, he worked on predictive and targeting algorithms, while at Samsung, he focused on optimizing search engines for insurance document retrieval.
<Experience Highlights>
[Reviewer / Program Committee] AAAI 2024-2025
[Reviewer] IEEE Transactions on Cognitive and Developmental Systems 2025
[Intership] Hyundai Motors Company, Tilda Corp., LG Telecom (a.k.a. LG Uplus), and Samsung Life Insurance
<Papers>
Jin Hyun Park., Representations learnt by SGD and Adaptive learning rules: Conditions that vary sparsity and selectivity in neural networks., 2022., [https://arxiv.org/pdf/2201.11653]
Seyyed Ali Ayati*, Jin Hyun Park*, Yichen Cai, Marcus Botacin., Making Acoustic Side-Channel Attacks on Noisy Keyboards Viable with LLM-Assisted Spectrograms' "Typo" Correction., 2025., [https://arxiv.org/pdf/2504.11622]
<Selected Projects>
Leveraging Depth and Attention Mechanisms for Improved RGB Image Inpainting (Part of graduate course: Computational Photography)
Developed a dual-encoder inpainting framework that fuses RGB and depth features with attention, markedly reducing artifacts and boosting SSD, PSNR, SSIM, and LPIPS over an RGB-only baseline.
Verified on both line- and square-mask tests; Grad-CAM shows the depth-aware model focuses on the missing regions and structural details better than the baseline, producing cleaner reconstructions.
Link: https://github.com/7201krap/CSCE748_Computational-Photography/
Personal Photo Search Engine (Part of graduate course: Information Storage and Retrieval)
Implemented image caption generation, query embedding, and advanced face recognition using pre-trained LLM/Transformer models.
Developed a personal photo search engine, integrating face-matching techniques and sentence-based semantic search features.
Confidence and Uncertainty Visualization in LLMs (Part of graduate course: Data Visualization)
Developed metrics to measure confidence and uncertainty in LLMs using probabilistic and statistical methods.
Created token-level visualizations, enabling a clear representation of confidence and uncertainty in the generated text.
Verified that the proposed metrics demonstrated a linear relationship with traditional evaluation metrics, such as BLEU, ROUGE, and METEOR
Link: https://github.com/7201krap/CSCE679_Data-Visualization/blob/main/Visualization_Project.pdf
<Topics>
Here are some topics that I prepared for you guys :) Ph.D. Qualifier Exam is highly recommended for your qualifier exam if you are preparing for it.
Email: jinhyun.park@tamu.edu
Linkedin: https://www.linkedin.com/in/jinhyunjasonpark/
Google Scholar: https://scholar.google.com/citations?user=D29N4akAAAAJ
Last update: Apr 2025