Kyung-Min Kim

/Users/kyungminkim/Desktop/kmkim.jpg

Biointelligence Lab (BI)
Department of Computer Science and Engineering
Seoul National University, 1 Gwanak-ro, Gwanak-gu
Seoul 151-744, Korea

·       Office : #314-1, Bldg. 302

·       Phone : +82-2-880-1847

·       Fax :     +82-2-875-2240

·       E-mail : kmkim (at) bi.snu.ac.kr

·       Homepage: http://bi.snu.ac.kr/~kmkim

·       Github: https://github.com/Kyung-Min

·       Google Scholar: https://scholar.google.co.kr/citations?user=LPNlQRUAAAAJ&hl=en

 

RESEARCH AREA 

 

·      Recommender System, Time-series Modeling, Computer Vision, Natural Language Processing, Video Understanding, Story Learning

 


WORK EXPERIENCES
 

 

·      NAVER CLOVA Gyeonggi, Korea (Aug 2018 ~ current)

·      Surromind Robotics, Seoul, Korea (April 2016 ~ July 2018)

 

 

TEACHING EXPERIENCES / INVITED TALKS (selected)

 

·      네이버 AI: 사람을 위한, 지속가능한, 글로벌을 지향하는 AI 플랫폼 그리고 미래@ AI Frontiers Summit Industrial Track, 한국통신학회, 2021

·      추천시스템 3.0: 딥러닝 시대에서 추천시스템이 가야할 방향 @ Korean Institute of Intelligent Systems 2021 (lecture)

·      How NAVER Develops Next Generation Recommender System @ ICEIC 2021 Industrial Session (invited talk)

·      Recommender system 3.0: Bias, Graph, and Causality for the Post Deep Learning Era (추천시스템 3.0: 딥러닝 후기시대에서 바이어스, 그래프, 그리고 인과관계의 중요성) @ DEVIEW 2020, LINE DevDay 2020

·      Deep Learning for Video Dialog @Naver AI Colloquim 2019 (invited talk)

·      Visual-Linguistic Representation for Video Question Answering @ACCV workshop 2018 (invited talk)

·      수술 보조자로서의 AI @대한종양외과학회 2018 (invited talk)

·      Visually-Grounded Question and Answering: from VisualQA to MovieQA @ICCE-Asia 2018 (tutorial)

·      TensorFlow 배워보는 Deep Learning 기초 @Facebook Innovation Laboratory 2018 (lecture)

·      Machine Learning & Deep Learning @LGU+ 2018 (lecture)

·      Machine Learning & Deep Learning @Kakao corporation 2017 (lecture)

·      Video Story QA Using Deep Learning: PororoQA and Deep Embedded Memory Networks @NAVER 2017 (invited talk)

 

 

INTERNATIONAL JOURNAL PUBLICATIONS 

 

·       D.-S. Hwang, J.-Y. Park, S.-Y. Kwon, K.-M. Kim, J.-W. Ha, and H.-W. Kim, “Self-supervised Auxiliary Learning for Graph Neural Networks via Meta-Learning”, IEEE Transactions on Pattern Analysis and Machine Intelligence (in revision)

·       H.-J. Park, S. Lee, D.-S. Hwang, J. Jeong, K.-M. Kim, J.-W. Ha, and H. J. Kim. “Learning Augmentation for GNNs with Consistency Regularization”, IEEE Access, 2021. (to appear)

·       K.-W. Park, J.-W. Ha, J. Lee, S. Kwon, K.-M. Kim, B.-T. Zhang, “M2FN: Multi-step Modality Fusion for Advertisement Image Assessment”, Applied Soft Computing Journal (SCI/SCIE), 103, May 2021, 107116.

 

 

INTERNATIONAL CONFERENCE PUBLICATIONS 

 

·       S.-J. Jung, Y.-J. Park, J. Jeong, K.-M. Kim, H. Kim, M.-K. Kim, H.-N. Kwak, Global-Local Item Embedding for Temporal Set Prediction, The ACM Conference Series on Recommender Systems (RecSys) LBR track, 2021.

·       K. -Y. Shin, H.-N. Kwak, K.-M. Kim, M.-K. Kim, Y.-J. Park, J. Jung, S.-J. Jung, One4all User Representation for Recommender Systems in E-commerce, arXiv preprint arXiv:2106.00573, 2021. [paper]

·       I.-J. Kwon, K.-M. Kim, J. Jeong, K.-Y. Shin, Y.-J. Park, B.-T. Zhang, AdamDGN: Adaptive Memory using Dynamic Graph Networks for Staleness Problem in Recommender System, 1th International Workshop on Online and Adaptive Recommender System at KDD, 2021 (spotlight).

·       (* equal contribution) S.-J. Jung*, K.-M. Kim*, H.-N. Kwak*, Y.-J. Park* "A Worrying Analysis of Probabilistic Time-series Models for Sales Forecasting". PMLR 137:98-105, 2020. [paper]

·       D.-S. Hwang, J.-Y. Park, S.-Y. Kwon, K.-M. Kim, J.-W. Ha, and H.-W. Kim, "Self-supervised Auxiliary Learning with Meta-paths for Heterogeneous Graphs". NeurIPS, 2020 (Acceptance Ratio 20.0%). [paper]

·       Y.-J Park, K. -Y. Shin, and K.-M. Kim, "Hop Sampling: A Simple Regularized Graph Learning for Non-Stationary Recommendation Environment". 16th International Workshop on Mining and Learning with Graphs on KDD, 2020. [paper]

·       K. -Y. Shin, Y.-J Park, and K.-M. Kim, "Multi-Manifold Learning for Large-scale Targeted Advertising System". 14th AdKDD workshop on KDD, 2020. [paper]

·       S.-G. Kim, K.-M. Kim, "Odds-Ratio Thompson Sampling to Adjust for Batch Effect", RecSys workshop on Bandit and Reinforcement Learning from User Interactions, 2020. [paper]

·       K.-M. Kim, D.-H. Kwak, H.-N. Kwak, Y.-J. Park, S.-K. Sim, J.-H. Cho, M.-K. Kim, J.-H. Kwon, Nako Sung and J.-W. Ha "Tripartite Heterogeneous Graph Propagation for Large-scale Social Recommendation ". The ACM Conference Series on Recommender Systems (RecSys) LBR track, 2019. (Acceptance Ratio 30.9%) [paper]

·       K.-M. Kim, S.-H. Choi, J.-H. Kim, and B.-T. Zhang, "Multimodal Dual Attention Memory for Video Story Question Answering". European Conference on Computer Vision (ECCV), 2018 (Acceptance Ratio 31.3%). [paper]

·       K.-M. Kim, M.-O. Heo, S.-H.Choi, and B.-T. Zhang, "DeepStory: Video Story QA by Deep Embedded Memory Networks", International Joint Conference on Artificial Intelligence (IJCAI), 2017. (Acceptance Ratio 25.9%, Travel Award) [paper] [code] [data]

* Invited paper to UAI 2017 MLTrain’s “Neural Abstract Machines” session

·       K.-M. Kim, S.-H. Choi, S.-J. Choi, S.-H. Kim, and B.-T. Zhang, "MuSM: Multimodal Sequence Memory for Video Story Question Answering". ICCV 2017 Workshop on The Joint Video and Language Understanding, 2017. (nominated as one of challenge winners) 

·       M.-O. Heo, K.-M. Kim, and B.-T. Zhang, "Story Learning from Kids Videos with Successive Event Order Embedding". ICCV 2017 Workshop on Closing the Loop Between Vision and Language, 2017.

·       K.-M. Kim, C.-J. Nan, M.-O. Heo, S.-H.Choi, and B.-T. Zhang, "PororoQA: A Cartoon Video Series Dataset for Story Understanding", NIPS 2016 Workshop on Large Scale Computer Vision System, 2016. [data]

·       K.-M. Kim, C.-J. Nan, M.-O. Heo, and B.-T. Zhang, "Pororobot: Child Tutoring Robot for English Education", International Symposium on Perception, Action, and Cognitive Systems (PACS), 2016. (Best Paper Award)

·       K.-M. Kim, C.-J. Nan, J.-W. Ha, Y.-J. Heo, and B.-T. Zhang, "Pororobot: A Deep Learning Robot That Plays Video Q&A Games", AAAI 2015 Fall Symposium on AI for Human-Robot Interaction (AI-HRI 2015), 2015. (Travel Award) [paper]

·       C.-J. Nan, K.-M. Kim, and B.-T. Zhang, "Social network analysis of TV drama characters via deep concept hierarchies", International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2015), 2015. (Oral presentation: Acceptance Ratio 18.0%) [paper]

·       J.-W. Ha, K.-M. Kim, and B.-T. Zhang, "Automated Visual-Lingusitc Knowledge Construction via Concept Learning from Cartoon Videos", In Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence (AAAI 2015), pp. 522-528, 2015. (Oral Presentation: Acceptance Ratio 11.9%, Featured in 10 Korea news articles(조선일보, 동아일보, 전자신문, 연합뉴스 ), 2 monthly magazines(파워코리아, 수학동아), YTN SCIENCE "사과나무", SBS "뉴스토리" ) [paper]

·       K.-M. Kim, J.-W. Ha, B.-J. Lee, and B.-T. Zhang, "Multimodal hierarchical models for visually grounded concept learning from cartoon videos", Asia-Pacific Conference on Computational Behavioral Sciences (APCCBS 2014), p. 38, 2014. (poster)

·       B.-J. Lee, J.-W. Ha, K.-M. Kim, and B. T. Zhang, "Evolutionary concept learning from cartoon videos by multimodal hypernetworks", IEEE Congress on Evolutionary Computation (CEC 2013), pp. 1186-1192, 2013. [paper]

 

DOMESTIC JOURNAL PUBLICATIONS 

 

·       M.-O. Heo, K.-M. Kim, B.-T. Zhang, "딥러닝 기반 비디오 스토리 학습 기술". Journal of Korea Multimedia Society, 20(3):23-40. 2016. 

·       C.-J. Nan, K.-M. Kim, B.-T. Zhang, "Social Network Analysis of TV Drama via Location Knowledge-learned Deep Hypernetworks", KIISE Transactions on Computing Practices, 22(11):612-617.2016. 

·       K.-M. Kim, J.-W. Ha, B.-T. Zhang, "Character-based Subtitle Generation by Learning of Multimodal Concept Hierarchy from Cartoon Videos" Journal of the Korea Information Science Society, 42(4):451-458, 2015

·       K.-M. Kim, H.-Y. Jang, and B.-T. Zhang, "Oversampling-based ensemble learning methods for imbalanced data", Journal of the Korean Institute of Information Scientists and Engineers: Computing Practices, 20(10):549-554, 2014

 

 

DOMESTIC CONFERENCE PUBLICATIONS (Not Updated Since 2016) 

 

·       C.-J. Nan, K.-M. Kim, B.-T. Zhang, "Social Network Analysis of TV Drama via Location Knowledge-learned Deep Hypernetworks", The 42th Winter Conference of Korean Institute of Information Scientists and Engineers (KIISE), pp. 822-824, 2015.

·       H.-Y. Jo, J.-H. Kim, S.-W, Yoon, K.-M. Kim, B.-T. Zhang, "Large-Scale Text Classification Methodology with Convolutional Neural Network", The 42th Winter Conference of Korean Institute of Information Scientists and Engineers (KIISE), pp. 725-727, 2015.

·       M.-O. Heo, D.-S. Han, K.-M. Kim, B.-T. Zhang, "A Deep Learning-based Story Learning Framework for Kids' Videos", The 42th Winter Conference of Korean Institute of Information Scientists and Engineers (KIISE), pp. 725-727, 2015.

·       K.-M. Kim, C.-J. Nan, J.-W. Ha, Y.-J. Heo, "Dual Deep Memories for Video Question Answering", The 42th Winter Conference of Korean Institute of Information Scientists and Engineers (KIISE), pp. 654-656, 2015.

·       Y.-J. Heo, K.-M. Kim, and B.-T. Zhang, "Pororobot: A Deep Learning Robot that Plays Video Q&A Games", The 42th Winter Conference of Korean Institute of Information Scientists and Engineers (KIISE), pp. 645-647, 2015 (Best Paper Award).

·       W.-Y. Kang, K.-M. Kim, C.-J. Nan, and B.-T. Zhang, "Visual Features Analysis for Deep Hypernetworks Learning Cartoon Videos", The 42th Winter Conference of Korean Institute of Information Scientists and Engineers (KIISE), pp. 630-632, 2015.

·       K.-M. Kim, J.-W. Ha, C.-J. Nan, B.-T. Zhang, "Learning and Inference in Cartoon Videos using Deep Hypernetworks Concept Structure", Korea Computer Congress 2015 (KCC 2015), pp. 814-816, 2015

·       K.-M. Kim, J.-W. Ha, C.-J. Nan, B.-T. Zhang, "Automated Visual-Linguistic Knowledge Construction via Deep Hypernetworks from Cartoon Videos", The Korean Society for Cognitive Science Annual Spring Conference, pp.72, 2015. (poster)

·       C.-J. Nan, K.-M. Kim, B.-T. Zhang, Social Network Analysis of TV Drama Characters via Deep Concept Hierarchies, Korea Computer Congress 2015 (KCC 2015), pp. 948-950, 2015

·       K.-M. Kim, J.-W. Ha, B.-T. Zhang, "Deep Hierarchical Networks for Learning Visually-Grounded Linguistic Concepts from Cartoon Videos", The 41th Winter Conference of Korean Institute of Information Scientists and Engineers (KIISE), pp. 518-520, 2014.

·       K.-M. Kim, J.-W. Ha, B. -J. Lee, B.-T. Zhang, "Character-based subtitle generation by learning of multimodal concept hierarchy from cartoon videos," Korea Computer Congress 2014 (KCC 2014), pp. 647-649. 2014. (Best Paper Award).

·       C.-J. Nan, K.-M. Kim, J.-W. Ha, B.-T. Zhang, "Hypernetwork-based concept network construction using convolutional neural networks", The 41th Winter Conference of Korean Institute of Information Scientists and Engineers (KIISE), pp. 586-588, 2014.

·       K.-W. On, B. -H. Kim, K.-M. Kim, D. -H. Kwak, T. -S. Park, B.-T. Zhang "CogDIEM: Database of Implicit Emotional Responses to Multimedia for Cognitive Computing", The 41th Winter Conference of Korean Institute of Information Scientists and Engineers (KIISE), pp. 571-573, 2014.

·       H. -Y. Jang, K.-M. Kim, J. -W. Park, S. -T. Hwang, B. -T. Zhang, "Efficient Ensemble Model Selection Using Bayesian Model Combination", The 40th Fall Conference of Korean Institute of Information Scientists and Engineers (KIISE), pp. 628-630, 2013. (Best Paper Award).

·       K.-M. Kim, H. -Y. Jang, J. -W. Park, S. -T. Hwang, B. -T. Zhang, "Oversampling-Based Ensemble Learning Methods for Imbalanced Data", The 40th Fall Conference of Korean Institute of Information Scientists and Engineers (KIISE), pp. 686-688, 2013. (Best Presentation Award)

·       K.-M. Kim, J.-W. Ha, B. -J. Lee, B.-T. Zhang, "Generating Image Descriptive Sentences via Multimodal Concept Networks and Language Models, Korea Computer Congress 2014 (KCC 2014), pp. 1538-1540, 2014.

·       K.-M. Kim, R. Ha, "In-Time Cache Eviction To Reduce Inefficient SSD Garbage Collection", Korea Computer Congress 2012 (KCC 2012), pp. 349-351, 2012. (Best Paper Award).

 

SELECTED AWARDS 

 

·       Challenge Winners of MovieQA Challenge for VideoQA Oct, 2017, University of Toronto & KIT & MIT

·       2017 Naver Ph.D. Fellowship

 

INTELLECTUAL PROPERTIES 

 

·      K.-M. Kim, M.-K. Kim, N.-H. Nak, S.-K. Sim. 개인화 컨텐츠 추천을 위한 실시간 그래프기반 임베딩 구축 방법 시스템 (METHOD AND SYSTEM OF REAL-TIME GRAPH-BASED EMBEDDING FOR PERSONALIZED CONTENT RECOMMENDATION). 2019-08-08. 10-2019-0096950.

·       K.–M. Kim,  H.-Y. Song, S.-Y. Kwon, N.-H. Nak. 클릭율 최대화를 위한 이미지 디자인 파라미터 최적화 방법 시스템 (METHOD AND SYSTEM FOR OPTIMIZING DESIGN PARAMETER OF IMAGE TO MAXIMIZE CLICK THROUGH RATE). 2019-03-07. 10-2019-0026343.

·       K.-M. Kim, B.-T. Zhang. 스토리 관련 질의 응답을 하는 심층 임베딩 메모리망. 2016-11-08. 등록번호 C-2016-026448ȣ.

·       B.-T. Zhang, K.-M. Kim, J.-H. Kim, H.-Y Jo, J-H. Chang, 문서 분류 장치 방법 (APPARATUS AND METHOD FOR CLASSIFYING DOCUMENT), 2016-08-19, 10-2016-0105071.

·       B.-T. Zhang, K.-M. Kim, J.-W. Ha, 멀티 모달리티 개념망 모델 구축 방법 멀티 모달리티 개념망 모델을 이용하는 모달리티 데이터 간의 복원 방법, 이를 수행하는 장치 (Method and system of Deep Concept Hierarchy for reconstruction of multi-modality data) , 10-2015-0053829, 2016-08-03. 등록번호 1646926

·       B.-T. Zhang, K.-M. Kim, T.-J. Kim, H.-Y. Jang, J.-W. Park, 손글씨 인식 방법 디바이스 (Method and device for recognizing handwriting), 2014-04-09, 10-2014-0042533.

 

EXHIBITIONS 

 

·       Pororobot in Roboworld 2016, Oct, 2016, KINTEX, Korea 

·       Pororobot in The 17th World Knowledge Forum (video presentation), Oct, 2016, Seoul, Korea