Ching-Yun (Irene) Ko @ IBM Research
 |
AI Scientist, |
About Me
I am a Research Scientist in the Trusted AI Group at the IBM Thomas J. Watson Research Center. Prior to joining IBM, I spent five wonderful years in the Department of Electrical Engineering and Computer Science at Massachusetts Institute of Technology working with Prof. Luca Daniel. I was also fortunate to have been part of the Department of Electrical and Electronic Engineering at the University of Hong Kong during my M.Phil. studies, where I grasped my first opportunity to conduct research in tensor decomposition arithmetic and system identification under the guidance of Prof. Ngai Wong and Prof. Kim Batselier (now at TU Delft). Earlier, I earned my B.Sc. degree from the School of Mathematics and Statistics at Wuhan University, where I spent four years doing nothing related to research but only drama, basketball, and maths.
My current research interests include next-generation language models, value alignment, representation learning, robustness and fairness analysis of AI models, tensor theory and applications, tensor-based numerical algorithms, and system identification. Visitors are more than welcome to reach out to me if you are interested in any of these topics!
Publications
Conference Papers
Ching-Yun Ko, Pin-Yu Chen, Payel Das, Yung-Sung Chuang, Luca Daniel. “On Robustness-Accuracy Characterization of Language Models using Synthetic Datasets,” in Proc. of the First Conference on Language Modeling (COLM’24). [PDF]
Ching-Yun Ko, Pin-Yu Chen, Payel Das, Jeet Mohapatra, Luca Daniel. "What Would Gauss Say About Representations? Probing Pretrained Image Models using Synthetic Gaussian Benchmarks, in Proc. of the 41th International Conference on Machine Learning (ICML’24). [PDF]
Peiqi Wang, Yingcheng Liu, Ching-Yun Ko, William M. Wells, Seth Berkowitz, Steven Horng, Polina Golland. “Sample-Specific Debiasing for Better Image-Text Models” in Proc. of the 8th Machine Learning for Healthcare Conference, PMLR 219:788-803, 2023. [PDF]
Yanwei Wang, Ching-Yun Ko, Pulkit Agrawal. “Visual pre-training for navigation: What can we learn from noise?,” in Proc. of 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS’23). [PDF]
Ching-Yun Ko, Jeet Mohapatra, Sijia Liu, Pin-Yu Chen, Luca Daniel, Lily Weng. “Revisiting contrastive learning through the lens of neighborhood component analysis: an integrated framework,” in Proc. of the 39th International Conference on Machine Learning (ICML’22). [PDF]
Wei Liao, Ching-Yun Ko, Tsui-Wei Weng, Luca Daniel, Joel Voldman. “Facile Prediction of Neutrophil Activation State from Microscopy Images: A New Dataset and Comparative Deep Learning Approaches,” in proc. of 2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI’22). [PDF]
Jeet Mohapatra*, Ching-Yun Ko*, Tsui-Wei Weng, Pin-Yu Chen, Sijia Liu, Luca Daniel, “Hidden Cost of Randomized Smoothing,” in Proc. of the 24th International Conference on Artificial Intelligence and Statistics (AISTATS’21). [PDF]
Jeet Mohapatra, Ching-Yun Ko, Tsui-Wei Weng, Pin-Yu Chen, Sijia Liu, Luca Daniel, “Higher-Order Certification For Randomized Smoothing,” in Proc. of the 34th Conference on Neural Information Processing Systems (NeurIPS’20). (Spotlight)
Rui Lin, Ching-Yun Ko, Zhuolun He, Cong Chen, Yuan Cheng, Hao Yu, Graziano Chesi, Ngai Wong, “HOTCAKE: Higher Order Tucker Articulated Kernels for Deeper CNN Compression,” in Proc. of the 15th International Conference on Solid-State & Integrated Circuit Technology (ICSICT’20). [PDF]
Zhaoyang Lyu*, Ching-Yun Ko*,Zhifeng Kong, Ngai Wong, Dahua Lin, Luca Daniel, “Fastened CROWN: Tightened Neural Network Robustness Certificates,” in Proc. of the 34th AAAI Conference on Artificial Intelligence (AAAI’20). [PDF] [Code]
Kim Batselier, Ching-Yun Ko, Ngai Wong, “Extended Kalman Filtering with Low-Rank Tensor Networks for MIMO Volterra System Identification,” in Proc. of the 58th Conference on Decision and Control (CDC’19). (Invited Paper). [PDF] [Code]
Ching-Yun Ko, Rui Lin, Shu Li, Ngai Wong, “MiSC: Mixed Strategies Crowdsourcing,” in Proc. of the 28th International Joint Conference on Artificial Intelligence (IJCAI’19). [PDF]
Cong Chen, Kim Batselier, Ching-Yun Ko, Ngai Wong, “A support tensor train machine,” in Proc. of the 28th International Joint Conference on Neural Networks (IJCNN’19).
Cong Chen, Kim Batselier, Ching-Yun Ko, Ngai Wong, “Matrix product operator restricted Boltzmann machines,” in Proc. International Joint Conference on Neural Networks (IJCNN’19).
Ching-Yun Ko*, Zhaoyang Lyu*, Tsui-Wei Weng, Luca Daniel, Ngai Wong, Dahua Lin, “POPCORN: Certifying robustness of recurrent neural networks,” in Proc. of the 36th International Conference on Machine Learning (ICML’19). [PDF] [Code]
Yuke Zhang, Ching-Yun Ko, Cong Chen, Kim Batselier, Ngai Wong, “Sparse Tensor Network System Identification for Nonlinear Circuit Macromodeling,” in Proc. of the 14th International Conference on Solid-State and Integrated Circuit Technology (ICSICT’18). (Invited Paper) [PDF]
Kim Batselier, Ching-Yun Ko, Anh-Huy Phan, Andrzej Cichocki, Ngai Wong, “Multilinear state space system identification with matrix product operators,” in Proc. of the 18th IFAC Symposium on System Identification (Sysid’18). [PDF]
Journal Articles
Ching-Yun Ko, Kim Batselier, Luca Daniel, Wenjian Yu, Ngai Wong, “Fast and Accurate Tensor Completion with Total Variation Regularized Tensor Trains,” IEEE Transactions on Image Processing, vol. 29, pp. 6918-6931, 2020. [PDF] [Code]
Ching-Yun Ko, Cong Chen, Zhuolun He, Yuke Zhang, Kim Batselier, Ngai Wong, “Deep Model Compression and Inference Speedup of Sum-Product Networks on Tensor Trains,” IEEE Transactions on Neural Networks and Learning Systems, vol. 31, no. 7, pp. 2665-2671, 2020. [PDF] [Code]
Kim Batselier, Ching-Yun Ko, Ngai Wong, “Tensor network subspace identification of polynomial state space models,” Automatica, vol. 95, pp. 87-196, 2018. [PDF] [Code]
Working Manuscript
Ching-Yun Ko*, Sihui Dai*, Payel Das*, Georgios Kollias, Subhajit Chaudhury, Aurelie Lozano, “MemReasoner: A Memory-augmented {LLM} Architecture for Multi-hop Reasoning,” presented at the First Workshop on System-2 Reasoning at Scale, NeurIPS’24. [short version]
Ching-Yun Ko, Pin-Yu Chen, Payel Das, Youssef Mroueh, Soham Dan, Georgios Kollias, Subhajit Chaudhury, Tejaswini Pedapati, Luca Daniel. “Large Language Models can be Strong Self-Detoxifiers.” arXiv preprint arXiv:2410.03818 (2024).
Kuo-Han Hung, Ching-Yun Ko, Ambrish Rawat, I. Chung, Winston H. Hsu, Pin-Yu Chen. “Attention Tracker: Detecting Prompt Injection Attacks in LLMs.” arXiv preprint arXiv:2411.00348 (2024).
Chung-Ting Tsai, Ching-Yun Ko, I. Chung, Yu-Chiang Frank Wang, Pin-Yu Chen. “Understanding and Improving Training-Free AI-Generated Image Detections with Vision Foundation Models.” arXiv preprint arXiv:2411.19117 (2024).
Professional Activities
Duties
Journal reviewer of Neurocomputing, IEEE Transactions on Image Processing, IEEE Transactions on Signal Processing, IEEE Transactions on Cybernetics, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, Optimization Methods and Software.
Conference reviewer of ICLR, NeurIPS, ICML, AISTATS, IJCAI, AAAI, CDC.
Program committee member of New Frontiers in Federated Learning: Privacy, Fairness, Robustness, Personalization and Data Ownership.
Talks
Invited speaker at UCLA Synthetic Data Workshop “Task-Agnostic Benchmarking of Pretrained Representations using Synthetic Data”, Apr. 2023.
Invited lightning talk in IJCAI 2019 workshop “Humanizing AI”, Aug. 2019.
Invited talk on “Tensor Decomposition in Machine Learning Applications” at Hangzhou Dianzi University, Dec. 2018.
Teaching
2.0966.7300(6.336)16.910 Introduction to Modeling and Simulation (graduate), Fall 2023.
MATH1853 Linear Algebra, Probability and Statistics (undergraduate), Fall 2018.
WHU Elite Program, Spring 2017, Fall 2016.
More About Me
I was (and am still!) a fan of basketball and drama during my undergrad fun time. Specially, I was the president of Shuying drama club in my sophomore year and the captain of women's basketball team of the school of Mathematics and Statistics in my junior year. I also enjoy travelling and hiking, especially hiking when travelling. I went on a trip for half month in the summer of 2017 and completed a 35-km hike with all the laugages. Later that year, I challanged myself with a 50-km hike (Hong Kong Trail) and won a place with my teammates. Research (My laziness) kept me occupied since then so no more fun stories (for now). XD
(Updated in 2025 Jan) In the whirlwind of AI’s rapid evolution, we may feel lost at times. Yet, it’s in embracing this journey with optimism, sharing our discoveries, and nurturing our well-being that we truly thrive.