Instance-Optimal Estimation with Multiple LLM Judges on a Budget
Evaluating large language models increasingly relies on LLM-as-a-judge protocols, but such evaluations remain costly: different judges have different prices and reliabilities, and …
Junghyun Lee
PhD Candidate in Artificial Intelligence
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.
Cumulative Distribution Regret Minimization with Max- Quantile Threshold in Multi-Armed Bandit
We study a new risk-averse bandit setting motivated by semiconductor manufacturing, where the quality of a recipe is judged not by its mean performance but by its weakest outcomes. …
jaeyoung-cha
Looking Through the Mirror: Minimax-Optimal Regularized Regrets in Online Learning and Bandits
We revisit regularized regret minimization under full-information and bandit feedback, where a learner optimizes an objective of the form $\langle r, \pi \rangle - \eta^{-1} …
Near-Optimal Clustering in Mixture of Markov Chains
We study the problem of clustering T trajectories of length H, each generated by one of K unknown ergodic Markov chains over a finite state space of size S. The goal is to …
GL-LowPopArt: A Nearly Instance-Wise Minimax-Optimal Estimator for Generalized Low-Rank Trace Regression
We present GL-LowPopArt, a novel Catoni-style estimator for generalized low-rank trace regression. Building on LowPopArt (Jang et al., 2024), it employs a two-stage approach -- …
TESSAR: Geometry-Aware Active Regression via Dynamic Voronoi Tessellation
Active learning improves training efficiency by selectively querying the most informative samples for labeling. While it naturally fits classification tasks–where informative …
seong-jin-cho
Provably Efficient Regularized Online RLHF with Generalized Bilinear Preferences
We consider the problem of *regularized* best-response max-regret minimization in online RLHF under general preferences and bandit feedback. While various regularizers are utilized …
A Jointly Efficient and Optimal Algorithm for Heteroskedastic Generalized Linear Bandits with Adversarial Corruptions
We consider the problem of heteroskedastic generalized linear bandits (GLBs) with adversarial corruptions, which subsumes various stochastic contextual bandit settings, including …
sanghwa-kim
Learning to Reason in LLMs by Expectation Maximization
Large language models (LLMs) solve reasoning problems by first generating a rationale and then answering. We formalize reasoning as a latent variable model and derive a …
Preliminary Empirical Study of Low-Rank, Hierarchical Gaussian Linear Bandits
Inspired by recent advances in multi-task bandits, we propose a new problem setting called low-rank, hierarchical Gaussian linear bandits, which combines low-rank structure with …