Junhoo Lee

I am a Ph.D. student at Seoul National University's Department of Intelligence and Information, under the guidance of Professor Nojun Kwak in the Machine Intelligence and Pattern Analysis Lab(MIPAL). I earned my bachelor's degree in Electronic and Computer Engineering(ECE) from the same institution in 2021.

I'm interested in computer vision, machine learning, meta learning, and dataset distillation. Much of my interest is currently focused on the dataset distillation and meta learning.

I am currently in search of an internship and welcome any potential opportunities. Please feel free to contact me if interested. Thank you! 🤝

Email  /  CV  / 

profile photo
Research
Any-Way meta learning
Junhoo Lee, Yearim Kim, Hyunho Lee, Nojun Kwak
AAAI 2024

In this paper, we present the discovery of 'label equivalence' and leverage it to develop the "anyway" learning paradigm, which transcends traditional fixed cardinality constraints in meta-learning.

SHOT: Suppressing the Hessian along the Optimization Trajectory
Junhoo Lee, Jayeon Yoo Nojun Kwak
NeurIPS 2023
project page / code / arXiv

We introduce linearity phenomenon in the inner loop in meta learning. Also, we introduce algorithm-architecture agnostic algorithm SHOT by utilizing this phenomenon.


Thanks for sharing the website template, Jon Barron. :)