Lune

Profiles2022-2026

Chulhee Yun

Kim Jaechul Graduate School of AI, KAIST

27Papers
23Last author
28Unique co-authors
1First author
0Solo

First and last author come from byline position, which some venues assign alphabetically.

Recent papers27

  1. Implicit Bias of Per-sample Adam on Separable Data: Departure from the Full-batch RegimeAuthor 3 of 3ICLR 2026 · 2 citations
  2. Minor First, Major Last: A Depth-Induced Implicit Bias of Sharpness-Aware MinimizationAuthor 3 of 3ICLR 2026 · 1 citation
  3. Implicit Bias and Loss of Plasticity in Matrix Completion: Depth Promotes Low-RanknessAuthor 2 of 2ICLR 2026
  4. Scaling Laws of SignSGD in Linear Regression: When Does It Outperform SGD?Author 3 of 3ICLR 2026
  5. Through the River: Understanding the Benefit of Schedule-Free Methods for Language Model TrainingAuthor 4 of 4NeurIPS 2025 · 9 citations
  6. From Linear to Nonlinear: Provable Weak-to-Strong Generalization through Feature LearningAuthor 3 of 3NeurIPS 2025 · 5 citations
  7. Arithmetic Transformers Can Length-Generalize in Both Operand Length and CountAuthor 4 of 4ICLR 2025
  8. Convergence and Implicit Bias of Gradient Descent on Continual Linear ClassificationAuthor 3 of 3ICLR 2025
  9. Incremental Gradient Descent with Small Epoch Counts is Surprisingly Slow on Ill-Conditioned ProblemsAuthor 3 of 3ICML 2025
  10. Lightweight Dataset Pruning without Full Training via Example Difficulty and Prediction UncertaintyAuthor 4 of 4ICML 2025
  11. Parameter Expanded Stochastic Gradient Markov Chain Monte CarloAuthor 3 of 5ICLR 2025
  12. Provable Benefit of Random Permutations over Uniform Sampling in Stochastic Coordinate DescentAuthor 3 of 3ICML 2025
  13. The Cost of Robustness: Tighter Bounds on Parameter Complexity for Robust Memorization in ReLU NetsAuthor 3 of 3NeurIPS 2025
  14. Understanding Sharpness Dynamics in NN Training with a Minimalist Example: The Effects of Dataset Difficulty, Depth, Stochasticity, and MoreAuthor 3 of 3ICML 2025
  15. Linear attention is (maybe) all you need (to understand Transformer optimization)Author 5 of 6ICLR 2024 · 92 citations
  16. Position Coupling: Improving Length Generalization of Arithmetic Transformers Using Task StructureAuthor 5 of 5NeurIPS 2024 · 28 citations
  17. Fundamental Benefit of Alternating Updates in Minimax OptimizationAuthor 3 of 3ICML 2024 · 14 citations
  18. Provable Benefit of Cutout and CutMix for Feature LearningAuthor 2 of 2NeurIPS 2024 · 11 citations
  19. DASH: Warm-Starting Neural Network Training in Stationary Settings without Loss of PlasticityAuthor 4 of 4NeurIPS 2024 · 8 citations
  20. Stochastic Extragradient with Flip-Flop Shuffling & Anchoring: Provable ImprovementsAuthor 2 of 3NeurIPS 2024 · 2 citations
  21. PLASTIC: Improving Input and Label Plasticity for Sample Efficient Reinforcement LearningAuthor 8 of 8NeurIPS 2023 · 50 citations
  22. Trajectory Alignment: Understanding the Edge of Stability Phenomenon via Bifurcation TheoryAuthor 2 of 2NeurIPS 2023 · 26 citations
  23. Fair Streaming Principal Component Analysis: Statistical and Algorithmic ViewpointAuthor 4 of 4NeurIPS 2023 · 11 citations
  24. On the Training Instability of Shuffling SGD with Batch NormalizationAuthor 2 of 2ICML 2023 · 6 citations
  25. Provable Benefit of Mixup for Finding Optimal Decision BoundariesAuthor 2 of 2ICML 2023 · 6 citations
  26. SGDA with shuffling: faster convergence for nonconvex-PŁ minimax optimizationAuthor 2 of 2ICLR 2023
  27. Minibatch vs Local SGD with Shuffling: Tight Convergence Bounds and BeyondAuthor 1 of 2ICLR 2022 · 47 citations
Dusk over the sea between two cliffs drawn in fine vertical lines