Publications
Conference Papers
[TOP-TIER] SwiftThief: Enhancing Query Efficiency of Model Stealing by Contrastive Learning, Jeonghyun Lee, Sungmin Han, Sangkyun Lee, IJCAI (the 33rd International Joint Conference on Artificial Intelligence), 2024 [pdf]
[TOP-TIER] Libra-CAM: An Activation-Based Attribution Based on the Linear Approximation of Deep Neural Nets and Threshold Calibration, Sangkyun Lee, Sungmin Han, IJCAI (the 31st International Joint Conference on Artificial Intelligence), 2022 (acceptance rate: 15%) [pdf]
Patent: 심층 신경망의 선형 근사를 기반으로 하는 활성화 기반 XAI 기법 및 임계 값 교정 방법 (An activation-based attribution based on the linear approximation of deep neural nets and threshold calibration), 출원번호: 10-2022-0088563
[TOP-TIER] Model Stealing Defense against Exploiting Information Leak Through the Interpretation of Deep Neural Nets, Jeonghyun Lee, Sungmin Han, Sangkyun Lee, IJCAI (the 31st International Joint Conference on Artificial Intelligence), 2022 (acceptance rate: 15%) [pdf]
Patent: 딥러닝 기반 분류 시스템에 대한 모델 탈취 방어 방법 (Model Stealing Defense for the Deep Learning based Classification System), 출원번호: 10-2022-0091273
Adversarial Attacks to Neural Networks on Manufacturing Product Image Data, Byeonggil Jung and Sangkyun Lee, CISC-W (Conference on Information Security and Cryptography-Winter), 2020 (best paper excellence award)
Enhancing the Adversarial Robustness of Compressed CNN with the Knowledge Distillation, Jeonghyun Lee and Sangkyun Lee, CISC-W (Conference on Information Security and Cryptography-Winter), 2020
The Vulnerability of Vehicle Speed Prediction 1D CNN Model Against Adversarial Attacks, Junhyung Kwon and Sangkyun Lee, CISC-W (Conference on Information Security and Cryptography-Winter), 2020
Machine Learning Data Poisoning Quantification using Linear Discriminant Analysis, Hyeongmin Cho and Sangkyun Lee, CISC-W (Conference on Information Security and Cryptography-Winter), 2020
Vulnerability of Federated Learning due to Malicious Attacks, Dongjun Min and Sangkyun Lee, CISC-W (Conference on Information Security and Cryptography-Winter), 2020
Sparse Portfolio Selection via the Sorted l1-Norm, Philipp J. Kremer, Sangkyun Lee, Malgorzata Bogdan, and Sandra Paterlini, EFMA (European Financial Management Association) Annual Meeting, 2018 [pdf]
Sparse Portfolio Selection via the Sorted L1 Norm, Malgorzata Bogdan, Philipp Kremer, Sandra Paterlini, and Sangkyun Lee, International Conference of the ERCIM WG on Computational and Methodological Statistics, 2018 [link]
[MAJOR] Fast Saddle-Point Algorithm for Generalized Dantzig Selector and FDR Control with the Ordered l1-Norm, Sangkyun Lee, Damian Brzyski and Malgorzata Bogdan, Proceedings of the 19th International Conference on Artificial Intelligence and Statistics (AISTATS), 2016 [pdf]
Kernel Completion for Learning Consensus Support Vector Machines in Bandwidth-Limited Sensor Networks, Sangkyun Lee and Christian Politz, International Conference on Pattern Recognition Applications and Methods (ICPRAM), 2014 [pdf] (best paper award nomination)
The Integer Approximation of Undirected Graphical Models, Nico Piatkowski, Sangkyun Lee, and Katharina Morik, International Conference on Pattern Recognition Applications and Methods (ICPRAM), 2014 [pdf]
Spatio-Temporal Random Fields: Compressible Representation and Distributed Estimation, Nico Piatkowski, Sangkyun Lee and Katharina Morik, European Conferences on Machine Learning (ECML), 2013 [pdf] (acceptance rate: 7%, best paper award in the journal track)
Separable Approximate Optimization of Support Vector Machines for Distributed Sensing, Sangkyun Lee, Marco Stolpe, and Katharina Morik, European Conferences on Machine Learning (ECML), 2012 [pdf]
ASSET: Approximate Stochastic Subgradient Estimation Training for Support Vector Machines, Sangkyun Lee and Stephen J. Wright, International Conference on Pattern Recognition Applications and Methods (ICPRAM), 2012 [pdf] [code]
[TOP-TIER] Manifold Identification of Dual Averaging Methods for Regularized Stochastic Online Learning, Sangkyun Lee and Stephen J. Wright, International Conference on Machine Learning (ICML), 2011. [pdf] [poster]
Decomposition Algorithms for Training Large-scale Semiparametric Support Vector Machines, Sangkyun Lee and Stephen J. Wright, European Conferences on Machine Learning (ECML), 2009. [pdf] [poster]
Journal Papers
[SCI: Q1] Similarity-Based Source Code Vulnerability Detection Leveraging Transformer Architecture: Harnessing Cross-Attention for Hierarchical Analysis, Sungmin Han, Miju Kim, Jaesik Kang, Kwangsoo Kim, Seungwoon Lee, Sangkyun Lee, IEEE Access, 2024 [pdf]
[SCI: Q1] CODE-SMASH: Source-Code Vulnerability Detection using Siamese and Multi-Level Neural Architecture, Sungmin Han, Hyunkyung Nam, Jaesik Kang, Kwangsoo Kim, Seungjae Cho, Sangkyun Lee, IEEE Access, 2024 [pdf]
Development of a Deep Learning-based Midterm PM2.5 Prediction Model Adapting to Trend Changes, Dong Jun Min, Hyerim Kim, Sangkyun Lee, The Transactions of the Korea Information Processing Society, 2024 [pdf]
[SCI: Q1 (TOP 7%)] Anomaly Candidate Extraction and Detection for Automatic Quality Inspection of Metal Casting Products using High-Resolution Images, Byeonggil Jung, Heegon You, Sangkyun Lee, Journal of Manufacturing Systems, 2023 [pdf].
[SCI: Q2] Activation Fine-Tuning of Convolutional Neural Networks for Improved Input Attribution Based on Class Activation Maps, Sungmin Han, Jeonghyun Lee, Sangkyun Lee, Applied Sciences, 2022 [pdf]
[SCI: Q2] Foreword to the special issue on advances in secure AI: Technology and applications, Sangkyun Lee, Applied Sciences, 2022 [pdf]
Abnormal Data Augmentation Method Using Perturbation Based on Hypersphere for Semi-Supervised Anomaly Detection, Byeonggil Jung , Junhyung Kwon, Dongjun Min , Sangkyun Lee, Journal of Korea Institute of Information Security & Cryptology, 2022 [pdf]
Deobfuscation Processing and Deep Learning-Based Detection Method for PowerShell-Based Malware, Ho-jin Jung , Hyo-gon Ryu, Kyu-whan Jo, Sangkyun Lee, Journal of Korea Institute of Information Security & Cryptology, 2022 [pdf]
Anomaly Detection in Multi-Host Environment Based on Federated Hypersphere Classifier, Junhyung Kwon, Byeonggil Jung, Hyungil Lee, and Sangkyun Lee, Electronics, 2022 [pdf]
Improving the Robustness of Model Compression by On-Manifold Adversarial Training , Junhyung Kwon and Sangkyun Lee, Future Internet, 2021 [pdf]
[SCI: Q2] Hunt for Unseen Intrusion: Multi-Head Self-Attention Neural Detector, Seongyun Seo, Sungmin Han, Janghyeon Park, Shinwoo Shim, Han-Eul Ryu, Byoungmo Cho, and Sangkyun Lee, IEEE Access, 2021 [pdf]
[SCI: Q2] Robust CNN Compression Framework for Security-Sensitive Embedded Systems, Jeonghyun Lee and Sangkyun Lee, Applied Sciences, 2021 [pdf]
[SCI: Q2] Data Quality Measures and Efficient Evaluation Algorithms for Large-Scale High-Dimensional Data, Hyeongmin Cho and Sangkyun Lee, Applied Sciences, 2021 [pdf]
[SSCI: Q1] Sparse Portfolio Selection via the sorted ℓ1 - Norm, Philipp J. Kremer*, Sangkyun Lee*, Małgorzata Bogdan, and Sandra Paterlini, Journal of Banking & Finance, 2020 [pdf]
[SCI: Q2] Structure Learning of Gaussian Markov Random Fields with False Discovery Rate Control, Sangkyun Lee, Piotr Sobczyk and Malgorzata Bogdan, Symmetry, 2019 [pdf]
[SCI: Q2] Compressed Learning of Deep Neural Networks for OpenCL-Capable Embedded Systems, Sangkyun Lee and Jeonghyun Lee, Applied Sciences, 2019 [pdf]
The mutational landscape of MYCN, Lin28b and ALKF1174L driven murine neuroblastoma mimics human disease, Bram De Wilde, Anneleen Beckers, Sven Lindner, Althoff Kristina, Katleen De Preter, Pauline Depuydt, Pieter Mestdagh, Tom Sante, Steve Lefever, Falk Hertwig, Zhiyu Peng, Le-ming Shi, Sangkyun Lee, Elien Vandermarliere, Lennart Martens, Björn Menten, Alexander Schramm, Matthias Fischer, Johannes Schulte, Jo Vandesompele and Frank Speleman, Oncotarget, 2017 [pdf]
[SCI: Q1] Integer undirected graphical models for resource-constrained systems, Nico Piatkowski, Sangkyun Lee and Katharina Morik, Neurocomputing, 2016 [pdf]
[SCI: Q1 (TOP 1%)] Mutational dynamics between primary and relapse neuroblastomas, Alexander Schramm, Johannes Koster, Yassen Assenov, Kristina Althoff, Martin Peifer, Ellen Mahlow, Andrea Odersky, Daniela Beisser, Corinna Ernst, Anton G. Henssen, Harald Stephan, Christopher Schroder, Lukas Heukamp, Anne Engesser, Yvonne Kahlert, Jessica Theissen, Barbara Hero, Frederik Roels, Janine Altmuller, Peter Nurnberg, Kathy Astrahantseff, Christian Gloeckner, Katleen De Preter, Christoph Plass, Sangkyun Lee, Holger N. Lode, Kai-Oliver Henrich, Moritz Gartlgruber, Frank Speleman, Peter Schmezer, Frank Westermann, Sven Rahmann, Matthias Fischer, Angelika Eggert, and Johannes H. Schulte, Nature Genetics, 2015 [pdf]
Sensitivity to cdk1-inhibition is modulated by p53 status in preclinical models of embryonal tumors, Melanie Schwermer, Sangkyun Lee, Johannes Koster, Tom van Maerken, Harald Stephan, Angelika Eggert, Katharina Morik, Johannes H. Schulte, and Alexander Schramm. Oncotarget, 2015 [pdf]
[SCI: Q1] Robust selection of cancer survival signatures from high-throughput genomic data using two-fold subsampling, Sangkyun Lee, Jorg Rahnenfuhrer, Michel Lang, Katleen De Preter, Pieter Mestdagh, Jan Koster, Rogier Versteeg, Raymond L. Stallings, Luigi Varesio, Shahab Asgharzadeh, Johannes H. Schulte, Kathrin Fielitz, Melanie Schwermer, Katharina Morik, and Alexander Schramm, PLOS ONE, 2014 [pdf]
[SCI: Q2] Spatio-Temporal Random Fields: Compressible Representation and Distributed Estimation, Nico Piatkowski, Sangkyun Lee and Katharina Morik, Machine Learning, 2013 [pdf]
[SCI: Q1] Manifold Identification in Dual Averaging Methods for Regularized Stochastic Online Learning, Sangkyun Lee and Stephen J. Wright, Journal of Machine Learning Research, 2012 [pdf]
Contributed
[특집원고] 쿼리 기반 복제 공격에 강건한 인공지능 모델 연구, 정보과학회지 (Communications of KIISE), vol. 40, no.11, pp.23~29, 2022.11. [pdf]
Workshop papers / Conference Talks
Spatio-Temporal Models For Sustainability, Nico Piatkowski, Sangkyun Lee and Katharina Morik, SustKDD Workshop in ACM Conference on Knowledge Discovery and Data Mining (KDD), 2012. [pdf]
Scalable Sochastic Gradient Descent with Improved Confidence, Sangkyun Lee and Christian Bockermann, Big Learning workshop in Neural Information Processing Systems (NIPS), 2012. [pdf]
Signal Processing Algorithms on Graphical Processing Units, Sangkyun Lee and Stephen J. Wright, INFORMS Annual Meeting, Invited Talk, 2009. [slides]
Decomposition and Stochastic Subgradient Algorithms for Support Vector Machines, Sangkyun Lee and Stephen J. Wright, International Symposium on Mathematical Programming (ISMP), 2009. [slides]
Book Chapters
Sparse Inverse Covariance Estimation for Graph Representation of Feature Structure, Sangkyun Lee, Interactive Knowledge Discovery and Data Mining in Biomedical Informatics, Springer, 2014. [pdf]
Stochastic Subgradient Estimation Training for Support Vector Machines, Sangkyun Lee and Stephen J. Wright, Mathematical Methodologies in Pattern Recognition and Machine Learning, 2013. [pdf]
Combining Information-based Supervised and Unsupervised Feature Selection, Sangkyun Lee, S.-J. Lee and B.-T. Zhang, Feature Extraction: Foundations and Applications, Springer, 2006. [Amazon]
Technical Reports
Approximate Stochastic Subgradient Estimation Training for Support Vector Machines, Sangkyun Lee and Stephen J. Wright, Nov 2011. [arxiv.org]
Manifold Identification in Dual Averaging Methods for Regularized Stochastic Online Learning, Sangkyun Lee and Stephen J. Wright, University of Wisconsin-Madison, July 2011 (supercedes April 2011 ver.). [pdf]
Implementing Algorithms for Signal and Image Reconstruction on Graphical Processing Units, Sangkyun Lee and Stephen J. Wright, University of Wisconsin-Madison, 2008. [pdf] [code]
Ph.D. Thesis
Optimization Methods for Regularized Convex Formulations in Machine Learning, Sangkyun Lee, University of Wisconsin-Maidson, 2011. [pdf]