Welcome to the Foundations of ML Lab! We're still at an early stage but research is being done actively. Please check back often for more updates.
We are focused on creating and applying machine learning algorithms for various fields, including NLP, vision, engineering, and bioinformatics.
Feel free to send us an email if you're interested in our work.
We like to touch on all aspects of machine learning, but our focus is primarily on representation learning. The following are some, but not all, of the central questions we aim to ask and answer in this lab.
How can we generate 'better' representation of objects?
What are some theoretical properties of existing or emerging representation learnings?
How will such new representations benefit other areas (like computer vision, robotics, healthcare, etc.)?
The recent list of topics we are tackling right now are:
Developing new types of representation.
How to 'reconcile' multiple different objects or concepts into a single representation.
How to efficiently learn new representations, or fine-tune existing representations.
In addition, we are actively collaborating with other labs, such as the Spatial Intelligence Lab, and the Brain-Inspired Computing Lab to explore beneficial application of ML in practice.
Prof. Byungkon Kang
Daye Eun - Accessible and fair ML
Jaeyong Jung
Sojeong Shin - LLM, Vision-based exercise monitoring
Prof. Francois Rameau - Computer vision
Prof. Yoon Seok Yang - Neuromorphic computing
Joon Han - Ph.D. student jointly advised with Prof. Simon Woo. Multi-modal intelligent user authentication
Hyewon Jung, MS (AMS) - Research staff member @Hyundai Motors
Hyo Jong Chung (Jointly advised by Prof. Yoon Seok Yang), BS - Currently MS student @ UT Austin, Dept. of CS
Ha Young Kim, MS
Daye Eun, MS - Currently in our PhD program!
Kyuri Kyeong, BS - Software engineer @JP Morgan
Si Yong Lee, Ryangjin Lee, Hawon Park, Yoora Kim, Byungkon Kang, and Yoon Seok Yang. "SpikeMamba: Spike-Driven State Space Models for Energy-Efficient Biomedical Sequence Modeling". In proceedings of MICCAI 2026.
Oleksii Nasypanyi, Jaemin Cho, Utku Ozbulak, Byungkon Kang, and François Rameau. "Seeing Through the Weights: Privacy Leakage in Scene Coordinate Regression". In proceedings of ECCV 2026. Project page (Courtesy of Jaemin Cho) Long oral presentation (~2%)
Juyoung Yun, Sol Choi, François Rameau, Byungkon Kang, and Zhoulai Fu. "Revisiting 16-bit Neural Network Training: A Practical Approach for Resource-Limited Learning". In proceedings of ICONIP 2025. Best Paper Award!
Joonkyu Han, Dennis Wong, Zhoulai Fu, and Byungkon Kang. "AuthZit: Personalized Visual-Spatial and Loci-Tagging Fallback Authentication", In proceedings of IEEE PRDC 2024.
Daye Eun and Byungkon Kang. "Accurate Embedding-based Log Determinant Optimization", In proceedings of ACM CIKM 2024.
Joonkyu Han, Dennis Wong, and Byungkon Kang. "AuthZit: Multi-Modal Authentication with Visual-Spatial and Text Secrets", In proceedings of ACNS 2023 (Poster).
Jeongmin Yoo, Yeeun Sohn, Yooha Bae, and Byungkon Kang. "Clean Bottle: A reverse vending machine that detects clean PET bottles with machine learning", In proceedings of ICIAE 2022.
Joonkyu Han, Byungkon Kang, and Dennis Wong. "HWAuth: Handwriting-Based Socially-Inclusive Authentication", In proceedings of SIGGRAPH Asia 2021 (Poster).
Hyo Jong Chung, Byungkon Kang, and Yoon Seok Yang. " N-DriverMotion: Driver motion learning and prediction using an event-based camera and directly trained spiking neural networks on Loihi 2", IEEE Open Journal of Vehicular Technology, 2024.
Hyunwook Koh, Won Gu, Hyo Jung Jang, Byungho Lee, and Byungkon Kang. "MiSurv: An integrative web cloud platform for user-friendly microbiome data analysis with survival responses", Microbiology Spectrum, 2023.
Hyojung Jang, Hyunwook Koh, Won Gu, and Byungkon Kang. “Integrative web cloud computing and analytics using MiPair for design-based comparative analysis with paired microbiome data”, Scientific Reports, 2022.
Won Gu, Jeongsup Moon, Crispen Chisina, Byungkon Kang, Taesung Park, and Hyunwook Koh. "MiCloud: A unified web platform for comprehensive microbiome data analysis", PLoS ONE, 2022.
Namhyuk Ahn, Byungkon Kang, Kyung-Ah Sohn, "Efficient Deep Neural Network for Photo-realistic Image Super-resolution", Pattern Recognition, 2022.
Byungkon Kang, Jisang Yoon, Ha Young Kim, Sung Jin Jo, Yourim Lee, Hye Jin Kam, "Deep-learning based automated terminology mapping in OMOP-CDM", Journal of the American Medical Informatics Association, 2021
Innovative human resource development for local intellectualization (IITP) - 40M KRW / year. 2023 - 2030
First career research grant (NRF) "Machine learning-based common factor extraction" - 30M KRW / year. 2020 - 2023 (completed)
Industry-Education-Research cooperation grant (Incheon TechnoPark) "Machine learning-based healthcare applications" - 6M KRW / 3 months, 2022 (completed)
Industry-Education-Research cooperation grant (Incheon TechnoPark) "Deep learning-based SNS data classification" - 12M KRW / 6 months, 2020 (completed)
Research fellow grant (NRF) "Error- and Ambiguity-robust Natural Language Understanding" - 50M KRW / year, 2016 - 2019 (completed)
Those who are seeking a research collaboration, please note that the funding for MS RA-ship is limited. PhD students have absolute priority over MS students in terms of financial support.
Possible areas of research include, but are not limited to: machine learning, natural language processing, and vision.
Successful candidates are expected to:
Be well-versed in (or at least willing to learn) mathematics and programming,
Conduct individual/group projects resulting in publishable research papers, and
Participate in periodic group discussions pertaining to on-going researches.
Exceptional undergraduate interns are also welcome (see below).
If you're interested, please send me an email to arrange for a discussion.
** This offer is valid only to students of SUNY Korea. Prospective students should first be admitted to the university before contacting me. I do not accept solicitations for admission - I have neither the authority, nor the intention to admit students without proper credentials.
The following are the conditions that I require from research participants from now on:
You MUST have basic knowledge of the topics that you wish to work on. It's normally advised that you complete either CSE352 or CSE353 with a reasonably good grade, but the exact requirements can be discussed. I can waive this requirement if you have graduate-level proficiency in mathematics and good programming skills.
If you're seeking credit for CSE487, then that CSE487 must be the only research course you're taking. I want you to stay focused.
Also with CSE487: You shouldn't be taking more than two (2) CS courses besides CSE487.
If you want to work as a paid research intern during the break, then I don't want you working elsewhere.
Don't get me wrong: Research is loads of fun. But I don't want you to confuse 'fun' with 'easy-going'.