Joon Kim

I am a junior at UC Berkeley majoring in EECS. My interest lies broadly in the complementary dynamics between Theoretical Computer Science and Computational Engineering and Sciences. I aspire to work in both directions, bringing the Computational Lens to other disciplines as well as being inspired by engineering and sciences for developing theoretical ideas in CS. My experience reflects this, as I have participated in multiple interdisciplinary research projects spanning Automotive Security, Healthcare, Artificial Intelligence, and Geography. Currently, I am independently studying various topics in mathematics and TCS to start getting involved in related research. In my free time, I usually practice playing the acoustic guitar, watch Netflix or Youtube, or pick up a random CS/Math textbook and hope I understand more than one sentence.

Detailed Summary

This summer, I participated in a SURF-REU program at University of Florida and researched automotive security, led by Professor Sandip Ray.
In Fall 24, I worked as an undergraduate researcher in the C.H.E.N. Lab, part of BAIR and CPH, led by Professor Irene Chen. There, I investigated the effect of zero-shot LLM inference on self-supervised learning alongside Ph.D. candidate Jichan Chung.
Previously, I worked as a research intern at JLK Inc., researching federated learning models applied to MRI image segmentation.
During my military service in 2023, I was fortunate to be advised by Professor Sejin Park to research defensive techniques against Deep Leakage in federated learning.
In Spring 22, I started my research journey with a URAP program in Studio.geo in the Geography department at Berkeley, led by Professor Clancy Wilmott. There, I explored uses and limitations of Generative AI models in generating ‘fake’ maps.

news

Aug 05, 2025	My first journal paper is published! Link
Aug 03, 2025	Successfully completed the SURF-REU program at University of Florida!
Dec 05, 2024	Participating in Advent of Code 2024. Repository can be found here.
Nov 25, 2024	Launching this personal homepage! Hooray!

latest posts

Dec 15, 2024	Advent of Code 2024 Week 2 Reflection
Dec 06, 2024	Advent of Code 2024 Week 1 Reflection
Nov 25, 2024	Hello!

selected publications

Intell.-Based Med.
In-silo federated learning vs. centralized learning for segmenting acute and chronic ischemic brain lesions

Joon Kim, Hoyeon Lee, Jonghyeok Park, and 7 more authors

Intelligence-Based Medicine, 2025

Abs DOI Bib

Objectives To investigate the efficacy of federated learning (FL) compared to industry-level centralized learning (CL) for segmenting acute infarct and white matter hyperintensity. Materials and methods This retrospective study included 13,546 diffusion-weighted images (DWI) from 10 hospitals and 8421 fluid-attenuated inversion recovery (FLAIR) images from 9 hospitals for acute (Task I) and chronic (Task II) lesion segmentation. We trained with datasets originated from 9 and 3 institutions for Task I and Task II, respectively, and externally tested them in datasets originated from 1 and 6 institutions each. For FL, the central server aggregated training results every four rounds with FedYogi (Task I) and FedAvg (Task II). A batch clipping strategy was tested for the FL models. Performances were evaluated with the Dice similarity coefficient (DSC). Results The mean ages (SD) for the training datasets were 68.1 (12.8) for Task I and 67.4 (13.0) for Task II. The frequency of male participants was 51.5 % and 60.4 %, respectively. In Task I, the FL model employing batch clipping trained for 360 epochs achieved a DSC of 0.754 ± 0.183, surpassing an equivalently trained CL model (DSC 0.691 ± 0.229; p < 0.001) and comparable to the best-performing CL model at 940 epochs (DSC 0.755 ± 0.207; p = 0.701). In Task II, no significant differences were observed amongst FL model with clipping, without clipping, and CL model after 48 epochs (DSCs of 0.761 ± 0.299, 0.751 ± 0.304, 0.744 ± 0.304). Few-shot FL showed significantly lower performance. Task II reduced training times with batch clipping (3.5–1.75 h). Conclusions Comparisons between CL and FL in identical settings suggest the feasibility of FL for medical image segmentation.
@article{KIM2025100283, title = {In-silo federated learning vs. centralized learning for segmenting acute and chronic ischemic brain lesions}, journal = {Intelligence-Based Medicine}, volume = {12}, pages = {100283}, year = {2025}, issn = {2666-5212}, doi = {https://doi.org/10.1016/j.ibmed.2025.100283}, url = {https://www.sciencedirect.com/science/article/pii/S2666521225000870}, author = {Kim, Joon and Lee, Hoyeon and Park, Jonghyeok and Park, Sang Hyun and Lee, Myungjae and Sunwoo, Leonard and Kim, Chi Kyung and Kim, Beom Joon and Kim, Dong-Eog and Ryu, Wi-Sun}, keywords = {Federated learning, Image segmentation, Ischemic brain lesion, Machine learning}, }

arxiv

Random Gradient Masking as a Defensive Measure to Deep Leakage in Federated Learning

Joon Kim, and Sejin Park

Preprint, 2024

DOI Bib

@article{kim2024randomgradientmaskingdefensive,
  title = {Random Gradient Masking as a Defensive Measure to Deep Leakage in Federated Learning},
  author = {Kim, Joon and Park, Sejin},
  year = {2024},
  eprint = {2408.08430},
  archiveprefix = {arXiv},
  primaryclass = {cs.LG},
  url = {https://arxiv.org/abs/2408.08430},
  doi = {10.48550/arXiv.2408.08430},
  journal = {Preprint},
}