KL
Kangwook Lee
cs.LGcs.AIcs.CLstat.MLcs.ITmath.ITcs.DCcs.CYcs.CRcs.CV
On Valency
published · living versionsW_2y3h6jgr·v1 · currentpublished
Where Should Learning Live? Regret Analysis for Text, Parameter, and Joint Adaptation
with Jungtaek Kim, Jaden Park, Jongwon Jeong, Emmanouil-Vasileios Vlatakis-Gkaragkounis
1 version
Preprints & journals
66 papers in the corpus · 2012–2025ReJump: A Tree-Jump Representation for Analyzing and Improving LLM Reasoning2512.00831v2 · Yuchen Zeng, Shuibai Zhang, Wonjun Kang et al.2025 · 0 citationsarXiv
ParallelBench: Understanding the Trade-offs of Parallel Decoding in Diffusion LLMs2510.04767v2 · Wonjun Kang, Kevin Galim, Seunghyuk Oh et al.2025 · 0 citationsarXiv
Draft-based Approximate Inference for LLMs2506.08373v3 · Kevin Galim, Ethan Ewer, Wonjun Kang et al.2025 · 0 citationsICLR 2026
In-Context Learning with Hypothesis-Class Guidance2502.19787v2 · Ziqian Lin, Shubham Kumar Bharti, Kangwook Lee2025 · 0 citationsarXiv
LLM-Lasso: A Robust Framework for Domain-Informed Feature Selection and Regularization2502.10648v3 · Erica Zhang, Ryunosuke Goto, Naomi Sagan et al.2025 · 2 citationsarXiv
VersaPRM: Multi-Domain Process Reward Model via Synthetic Reasoning Data2502.06737v2 · Thomas Zeng, Shuibai Zhang, Shutong Wu et al.2025 · 0 citationsarXiv
Self-Improving Transformers Overcome Easy-to-Hard and Length Generalization Challenges2502.01612v2 · Nayoung Lee, Ziyang Cai, Avi Schwarzschild et al.2025 · 1 citationarXiv
Task Vectors in In-Context Learning: Emergence, Formation, and Benefit2501.09240v1 · Liu Yang, Ziqian Lin, Kangwook Lee et al.2025 · 0 citationsarXiv
Multi-Bin Batching for Increasing LLM Inference Throughput2412.04504v2 · Ozgur Guldogan, Jackson Kunde, Kangwook Lee et al.2024 · 0 citationsarXiv
Forward and Inverse Simulation of Pseudo-Two-Dimensional Model of Lithium-Ion Batteries Using Neural Networks2412.13200v2 · Myeong-Su Lee, Jaemin Oh, Dong-Chan Lee et al.2024 · 14 citationsComputer Methods in Applied Mechanics and Engineering Volume 438, Part B, 1 April 2025, 117856
Rare-to-Frequent: Unlocking Compositional Generation Power of Diffusion Models on Rare Concepts with LLM Guidance2410.22376v4 · Dongmin Park, Sebin Kim, Taehong Moon et al.2024 · 0 citationsarXiv
Parameter-Efficient Fine-Tuning of State Space Models2410.09016v3 · Kevin Galim, Wonjun Kang, Yuchen Zeng et al.2024 · 1 citationarXiv
Everything Everywhere All at Once: LLMs can In-Context Learn Multiple Tasks in Superposition2410.05603v1 · Zheyang Xiong, Ziyang Cai, John Cooper et al.2024 · 0 citationsarXiv
ENTP: Encoder-only Next Token Prediction2410.01600v3 · Ethan Ewer, Daewon Chae, Thomas Zeng et al.2024 · 0 citationsarXiv
Looped Transformers for Length Generalization2409.15647v5 · Ying Fan, Yilun Du, Kannan Ramchandran et al.2024 · 0 citationsarXiv
Buffer-based Gradient Projection for Continual Federated Learning2409.01585v1 · Shenghong Dai, Jy-yong Sohn, Yicong Chen et al.2024 · 0 citationsarXiv
From Artificial Needles to Real Haystacks: Improving Retrieval Capabilities in LLMs by Finetuning on Synthetic Data2406.19292v2 · Zheyang Xiong, Vasilis Papageorgiou, Kangwook Lee et al.2024 · 1 citationarXiv
Dual Operating Modes of In-Context Learning2402.18819v2 · Ziqian Lin, Kangwook Lee2024 · 1 citationarXiv
Can Mamba Learn How to Learn? A Comparative Study on In-Context Learning Tasks2402.04248v2 · Jongho Park, Jaeseung Park, Zheyang Xiong et al.2024 · 16 citationsarXiv
Can MLLMs Perform Text-to-Image In-Context Learning?2402.01293v3 · Yuchen Zeng, Wonjun Kang, Yicong Chen et al.2024 · 1 citationarXiv
Hierarchical Deep Reinforcement Learning-Based Propofol Infusion Assistant Framework in Anesthesia.35853065 · Yun, Won Joon, Shin, Myungjae, Mohaisen, David et al.2024 · 17 citationsIEEE transactions on neural networks and learning systems. 2024;35(2):2510-2521
Looped Transformers are Better at Learning Learning Algorithms2311.12424v3 · Liu Yang, Kangwook Lee, Robert Nowak et al.2023 · 0 citationsarXiv
Super-Resolution Emulation of Large Cosmological Fields with a 3D Conditional Diffusion Model2311.05217v2 · Adam Rouhiainen, Michael Gira, Moritz Munchmeyer et al.2023 · 11 citationsarXiv
Image Clustering Conditioned on Text Criteria2310.18297v4 · Sehyun Kwon, Jaeseung Park, Minkyu Kim et al.2023 · 5 citationsarXiv
The Expressive Power of Low-Rank Adaptation2310.17513v3 · Yuchen Zeng, Kangwook Lee2023 · 9 citationsarXiv
Intuitive Access to Smartphone Settings Using Relevance Model Trained by Contrastive Learning2307.09177v1 · Joonyoung Kim, Kangwook Lee, Haebin Shin et al.2023 · 0 citationsarXiv
Mini-Batch Optimization of Contrastive Loss2307.05906v1 · Jaewoong Cho, Kartik Sreenivasan, Keon Lee et al.2023 · 2 citationsarXiv
Predictive Pipelined Decoding: A Compute-Latency Trade-off for Exact LLM Decoding2307.05908v2 · Seongjun Yang, Gibbeum Lee, Jaewoong Cho et al.2023 · 3 citationsarXiv
Teaching Arithmetic to Small Transformers2307.03381v1 · Nayoung Lee, Kartik Sreenivasan, Jason D. Lee et al.2023 · 8 citationsarXiv
Variation Spaces for Multi-Output Neural Networks: Insights on Multi-Task Learning and Network Compression2305.16534v3 · Joseph Shenouda, Rahul Parhi, Kangwook Lee et al.2023 · 1 citationJournal of Machine Learning Research, vol. 25, no. 231, pp. 1-40, 2024
Prompted LLMs as Chatbot Modules for Long Open-domain Conversation2305.04533v1 · Gibbeum Lee, Volker Hartmann, Jongho Park et al.2023 · 34 citationsarXiv
Improving Fair Training under Correlation Shifts2302.02323v1 · Yuji Roh, Kangwook Lee, Steven Euijong Whang et al.2023 · 1 citationarXiv
Optimizing DDPM Sampling with Shortcut Fine-Tuning2301.13362v4 · Ying Fan, Kangwook Lee2023 · 11 citationsarXiv
Looped Transformers as Programmable Computers2301.13196v1 · Angeliki Giannou, Shashank Rajput, Jy-yong Sohn et al.2023 · 5 citationsarXiv
Equal Improvability: A New Fairness Notion Considering the Long-term Impact2210.06732v2 · Ozgur Guldogan, Yuchen Zeng, Jy-yong Sohn et al.2022 · 2 citationsarXiv
Outlier-Robust Group Inference via Gradient Space Clustering2210.06759v1 · Yuchen Zeng, Kristjan Greenewald, Kangwook Lee et al.2022 · 0 citationsarXiv
PathProx: A Proximal Gradient Algorithm for Weight Decay Regularized Deep Neural Networks2210.03069v4 · Liu Yang, Jifan Zhang, Joseph Shenouda et al.2022 · 0 citationsarXiv
Utilizing Language-Image Pretraining for Efficient and Robust Bilingual Word Alignment2205.11616v2 · Tuan Dinh, Jy-yong Sohn, Shashank Rajput et al.2022 · 0 citationsarXiv
Breaking Fair Binary Classification with Optimal Flipping Attacks2204.05472v2 · Changhun Jo, Jy-yong Sohn, Kangwook Lee2022 · 5 citationsarXiv
GenLabel: Mixup Relabeling using Generative Models2201.02354v1 · Jy-yong Sohn, Liang Shang, Hongxu Chen et al.2022 · 4 citationsarXiv
Improved Input Reprogramming for GAN Conditioning2201.02692v3 · Tuan Dinh, Daewon Seo, Zhixu Du et al.2022 · 2 citationsarXiv
Augment & Valuate : A Data Enhancement Pipeline for Data-Centric AI2112.03837v1 · Youngjune Lee, Oh Joon Kwon, Haeju Lee et al.2021 · 7 citationsarXiv
Improving Fairness via Federated Learning2110.15545v3 · Yuchen Zeng, Hongxu Chen, Kangwook Lee2021 · 34 citationsarXiv
Gradient Inversion with Generative Image Prior2110.14962v1 · Jinwoo Jeon, Jaechang Kim, Kangwook Lee et al.2021 · 94 citationsarXiv
Sample Selection for Fair and Robust Training2110.14222v1 · Yuji Roh, Kangwook Lee, Steven Euijong Whang et al.2021 · 25 citationsarXiv
Coded-InvNet for Resilient Prediction Serving Systems2106.06445v2 · Tuan Dinh, Kangwook Lee2021 · 2 citationsarXiv
Permutation-Based SGD: Is Random Optimal?2102.09718v2 · Shashank Rajput, Kangwook Lee, Dimitris Papailiopoulos2021 · 1 citationarXiv
FairBatch: Batch Selection for Model Fairness2012.01696v2 · Yuji Roh, Kangwook Lee, Steven Euijong Whang et al.2020 · 56 citationsarXiv
SLM: Learning a Discourse Language Representation with Sentence Unshuffling2010.16249v1 · Haejun Lee, Drew A. Hudson, Kangwook Lee et al.2020 · 42 citationsarXiv
Accordion: Adaptive Gradient Communication via Critical Learning Regime Identification2010.16248v1 · Saurabh Agarwal, Hongyi Wang, Kangwook Lee et al.2020 · 14 citationsarXiv
Career total: 252 works. 66 are in this corpus.Showing the 50 most recent.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.