To extend the generality of AI, we've studied
learning-to-learn (a.k.a meta-learning or few-shot learning) [c2, p1, c3, c4],
unsupervised learning of compositional representation such as object-centric representation [c5, c6],
and agent learning incorporating episodic memory [w1].
I am interested in compositional representation which can be generally utilized for a large range of tasks like human intelligence.
(c: peer-reviewed conference, w: peer-reviewed workshop, p: preprint, * = equal contributions)
J. Yoon*, T. Kim*, O. Dia, S. Kim, Y. Bengio, S. Ahn
Spotlight Presentation (top 3.5% = 168/4856)
[c1] A hierarchical-structured dictionary learning for image classification (arXiv)
J. Yoon, J. Choi, CD Yoo
International Conference on Image Processing (ICIP) 2014
Have serviced as a reviewer from 2021 for NeurIPS, ICML, and ICLR.