Junghyun Lee
PhD Candidate in AI
Kim Jaechul Graduate School of AI, KAIST
PhD candidate at KAIST AI, jointly advised by Se-Young Yun and Chulhee Yun. I work on interactive machine learning, theoretical aspects of LLMs, learning/optimization theory, and statistical analysis of large networks.
Publications 29
Instance-Optimal Estimation with Multiple LLM Judges on a Budget New
NeurIPS 2026
ICML 2026 CTB Workshop
· Best Paper Award (Long Paper track)
Looking Through the Mirror: Minimax-Optimal Regularized Regrets in Online Learning and Bandits New
NeurIPS 2026
CKAIA 2026
· Distinguished Paper Award
Provably Efficient Regularized Online RLHF with Generalized Bilinear Preferences New
NeurIPS 2026
ICML 2026 Pluralistic Alignment Workshop
CKAIA 2026
· Distinguished Paper Award
GL-LowPopArt: A Nearly Instance-Wise Minimax-Optimal Estimator for Generalized Low-Rank Trace Regression
AISTATS 2026
Near-Optimal Clustering in Mixture of Markov Chains
AISTATS 2026
TESSAR: Geometry-Aware Active Regression via Dynamic Voronoi Tessellation
ICLR 2026
Pointwise or Pairwise: When Do Pairwise Losses Help Reward Learning, Provably?
arXiv:2609.37209
A Jointly Efficient and Optimal Algorithm for Heteroskedastic Generalized Linear Bandits with Adversarial Corruptions
arXiv:2602.10971
Cumulative Distribution Regret Minimization with Max-Quantile Threshold in Multi-Armed Bandit
KCC 2026
· Best Paper Award
AdaSTaR: Adaptive Data Sampling for Training Self-Taught Reasoners
NeurIPS 2025
FlickerFusion: Intra-trajectory Domain Generalizing Multi-Agent RL
ICLR 2025
NeurIPS 2024 OWA Workshop
Probability-Flow ODE in Infinite-Dimensional Function Spaces
ICLR 2025 DeLTa Workshop
arXiv:2503.10219
Learning to Reason in LLMs by Expectation Maximization
arXiv:2512.20169
Preliminary Empirical Study of Low-Rank, Hierarchical Gaussian Linear Bandits
KSC 2025
A Unified Confidence Sequence for Generalized Linear Models, with Applications to Bandits
NeurIPS 2024
ICML 2024 ARLET Workshop
· Oral
CKAIA 2024
· Best Paper Award
Gradient Descent with Polyak's Momentum Finds Flatter Minima via Large Catapults
ICML 2024 HiLD Workshop
NeurIPS 2023 M3L Workshop
· Oral
arXiv:2311.15051
Querying Easily Flip-flopped Samples for Deep Active Learning
ICLR 2024
Improved Regret Bounds of (Multinomial) Logistic Bandits via Regret-to-Confidence-Set Conversion
AISTATS 2024
On the Estimation of Linear Softmax Parametrized Markov Chains
KCC 2024
Empirical Analyses of Corruption in the Clustering of Block MDPs
KIISE Transactions on Computing Practices, vol. 30, no. 4, pp. 187–192 (2024)
Fair Streaming Principal Component Analysis: Statistical and Algorithmic Viewpoint
NeurIPS 2023
Flooding with Absorption: An Efficient Protocol for Heterogeneous Bandits over Complex Networks
OPODIS 2023
· Best Student Paper Award
Nearly Optimal Latent State Decoding in Block MDPs
AISTATS 2023
CKAIA 2022
· Best Paper Award
On the Estimation of Linear Softmax Parametrized Probability Distributions
KSC 2023
Fast and Efficient MMD-based Fair PCA via Optimization over Stiefel Manifold
AAAI 2022
JKAIA 2021
· Best Paper Award
Preliminary Empirical Analyses of Clustering in Block MDPs
KSC 2022
· Distinguished Presentation Award
A Statistical Analysis of Stochastic Gradient Noises for GNNs
KCC 2022
Preliminary Evaluation of SWAY in Permutation Decision Space via a Novel Euclidean Embedding
SSBSE 2021
Some Loci in the Animation of a Sangaku Diagram
Forum Geometricorum, vol. 16, pp. 187–191 (2016)