My research objective is modeling complex interactions in high dimensional large-scale time series from various domains. Main research interest is large-scale inference with probabilistic graphical model. Nowadays, I am trying to solve interaction related problems with various Machine Learning algorithms such as Topic Models, Neural Networks for timeseries and various point processes.
The research topic which I have been fascinated is automation of machine learning. Human machine learning experts have led the success of machine learning, applied in various cutting-edge technology industries, in recent years. However the tremendous attention from industries and also academia was not able to make non-experts apply off-the-shelf machine learning algorithms without domain knowledge, although approachable methods through open source packages have been developed. In this context, Bayesian optimization can be used to optimize various types of parameters. Therefore I think it will be a potent strategy for finding the extrema of objective functions that are expensive to evaluate.
I am interested in interpretable machine learning. When machine learning is introduced to the real world problem, the interpretability of the model plays an important role in determining whether the result of the model is reliable. In the case of a general deep neural network model, it is impossible to know which process has been used to derive a specific result, so that even if a wrong decision is made, it is impossible to know whether the result is wrong before the problem occurs. In particular, the interpretability of the model is even more important in order to introduce machine learning into a problem where a misconducted situation causes a very large problem, both socially and morally.
I am interested in meta-learning, a line of research that aims to learn certain components of learning algorithms themselves. More specifically, I am investigating new ways to incorporate meta-learned knowledge into the process of training deep neural networks. We expect such learned learning algorithms to be more efficient compared to traditional hand-designed algorithms.
- Nayeong Kim - Spring 2017
- Byungjin Park - Spring 2017
- Seung Ho Lee - Spring 2017
- Byunghun So - Spring 2017
- Nayeong Kim - Fall 2016
- Byungjin Park - Fall 2016
- Seung Ho Lee - Fall 2016
- Woo Chang Jeong - Fall 2016
- Jaehan Park - Summer 2016
- Byungjin Park - Spring 2016
- SunUng Mun - Spring 2016
- Nayeong Kim - Spring 2016
- Minho Kim - Fall 2015
- Woo Hyeon Shim - Summer 2015, Fall 2015
Alumni / Alumnae
- Saehoon Kim,
"Binary Embedding: Theory and Applications,"
Ph.D., February 2018 (now in AItrics).
- Juho Lee,
"Efficient Bayesian Nonparametric Inference: Tree-Based Methods and Power-Law Models,"
Ph.D., February 2018 (now in University of Oxford).
- Bonkon Koo,
"Brain-to-brain interface for animal control,"
Ph.D., August 2017 (now in Samsung Electronics).
- Yong-Deok Kim,
"Bayesian learning for collaborative prediction,"
Ph.D., February 2015 (now in Samsung Electronics).
- Hyohyeong Kang,
"Bayesian common spatial patterns for EEG classification,"
Ph.D., February 2015 (now in Samsung Electronics).
- Kye-Hyeon Kim,
"Learning with minimax paths on graphs,"
Ph.D., February 2015 (now in Intel Korea).
- Yunjun Nam,
"Tongue-machine interface with glossokinetic potentials,"
Ph.D., February 2014 (now in RIKEN, Japan).
- Yongsoo Kim,
"Probabilistic inference in context-specific dynamic networks,"
Ph.D., August 2013 (now in The Netherlands Cancer Institute, Netherlands).
- Sunho Park,
"Embeddings for multi-class and multi-label learning,"
Ph.D., February 2013 (now in University of Texas Southwestern Medical Center at Dallas, USA).
- Jiho Yoo,
"Learning with matrix co-factorization,"
Ph.D., February 2012 (now in Samsung Advanced Institute of Technology).
- Sangki Kim,
"Dynamic hand gesture recognition with accelerometer,"
Ph.D., February 2010 (now in Vuno).
- Jong Kyoung Kim,
"Probabilistic models for motif discovery in biopolymer sequences,"
Ph.D., February 2010 (now in DGIST).
- Hyekyoung Lee,
"Nonnegative matrix and tensor factorization methods for spectral EEG classification,"
Ph.D., February 2009 (now in Seoul National University Medical School).
- Jiyuu Yi, MS, February 2018 (now in AItrics)
- Inhyuk Jo, MS, February 2018 (now in AItrics)
- Minseop Park, MS, February 2018 (now in AItrics)
- Vu Thi Hanh, MS, August 2017
- Sojeong Ha, MS, February 2016 (now in Samsung Electronics).
- Hien Duy Pham, MS, August 2014 (now in Xeron Healthcare).
- Huong Thi Pham, MS, August 2014 (now in Xeron Healthcare).
- Eunsil Gim, MS, February 2013 (now in Samsung).
- Thu Hoai Tran, MS, February 2013 (now in Vietnam).
- Huyen Le Thanh, MS, August 2012 (now in Vietnam).
- Sangmin Lee, MS, February 2012 (now in Hyundai).
- Hyojung Shin, MS, August 2011 (now in Microsoft, Redmond).
- Shounan An, MS, February 2010 (now in LG).
- Sun Ho Lee, MS, August 2007 (now in Amazon.com).
- Jong Kyoung Kim, MS, February 2006 (co-supervised).
- Sunho Park, MS, February 2006.
- Minje Kim, MS, February 2006 (now in Indiana University).
- Jaehwan Kim, MS, February 2005 (co-supervised) (now in ETRI).
- Kijeong Nam, MS, August 2005 (now in University of Maryland, College Park).
- Heeyoul Choi, MS, February 2005 (now in Handong University)
- Sookjeong Kim, MS, February 2005 (now in Ulsan)
- Seong-Cheol Park, MS, February 2004 (co-supervised) (now in KT).
- Yong-Choon Cho, MS, February 2004 (co-supervised) (now in Samsung).
- Inseon Jang, MS, February 2004 (now in ETRI).
- Yongjin Lee, MS, February 2004 (now in University of Washington, Seattle).
- Hyejin Kim, MS, August 2003 (co-supervised) (now in ETRI).
- Hye-Kyoung Lee, MS, February 2003
- Sangki Kim, MS, February 2003 (co-supervised)
- Heonseok Hong, MS, February 2002 (now in Samsung).
- O Young Lee, MS, February 2001 (now in Hyundai).
- Youngki Lyu, MS, February 2000 (now in Samsung)
- Jeong-Min Yun
- Yoonseop Kang
- Jungsoo Ahn
- Heuna Kim
- Soyeon Lee
- Pilwon Kim