15 citations · 56 across the 16 of their papers we have counts for
17 papers
Training Small Language Models for Transferable Tool Selection
Zhaoyi Zhou, Akshayaa Magesh, Jeongyeol Kwon +4
As LLM agents operate over corpora of more than ten thousand tools, selecting the right tool for a complex multi-step user instruction becomes a central bottleneck. Tool descriptio…
Imbalanced Gradients in RL Post-Training of Multi-Task LLMs
Runzhe Wu, Ankur Samanta, Ayush Jain +7
Multi-task post-training of large language models (LLMs) is typically performed by mixing datasets from different tasks and optimizing them jointly. This approach implicitly assume…
A Classification View on Meta Learning Bandits
Mirco Mutti, Jeongyeol Kwon, Shie Mannor +1
Contextual multi-armed bandits are a popular choice to model sequential decision-making. E.g., in a healthcare application we may perform various tests to asses a patient condition…
RL in Latent MDPs is Tractable: Online Guarantees via Off-Policy Evaluation
Jeongyeol Kwon, Shie Mannor, Constantine Caramanis +1
In many real-world decision problems there is partially observed, hidden or latent information that remains fixed throughout an interaction. Such decision problems can be modeled a…
Prospective Side Information for Latent MDPs
Jeongyeol Kwon, Yonathan Efroni, Shie Mannor +1
In many interactive decision-making settings, there is latent and unobserved information that remains fixed. Consider, for example, a dialogue system, where complete information ab…
Feed Two Birds with One Scone: Exploiting Wild Data for Both Out-of-Distribution Generalization and Detection
Haoyue Bai, Gregory Canal, Xuefeng Du +3
Modern machine learning models deployed in the wild can encounter both covariate and semantic shifts, giving rise to the problems of out-of-distribution (OOD) generalization and OO…
A Fully First-Order Method for Stochastic Bilevel Optimization
Jeongyeol Kwon, Dohyun Kwon, Stephen Wright +1
We consider stochastic unconstrained bilevel optimization problems when only the first-order gradient oracles are available. While numerous optimization methods have been proposed…
Tractable Optimality in Episodic Latent MABs
Jeongyeol Kwon, Yonathan Efroni, Constantine Caramanis +1
We consider a multi-armed bandit problem with latent contexts, where an agent interacts with the environment for an episode of time steps. Depending on the length of the ep…
Reward-Mixing MDPs with a Few Latent Contexts are Learnable
Jeongyeol Kwon, Yonathan Efroni, Constantine Caramanis +1
We consider episodic reinforcement learning in reward-mixing Markov decision processes (RMMDPs): at the beginning of every episode nature randomly picks a latent reward model among…
Coordinated Attacks against Contextual Bandits: Fundamental Limits and Defense Mechanisms
Jeongyeol Kwon, Yonathan Efroni, Constantine Caramanis +1
Motivated by online recommendation systems, we propose the problem of finding the optimal policy in multitask contextual bandits when a small fraction of tasks (users) are…
Reinforcement Learning in Reward-Mixing MDPs
Jeongyeol Kwon, Yonathan Efroni, Constantine Caramanis +1
Learning a near optimal policy in a partially observable system remains an elusive challenge in contemporary reinforcement learning. In this work, we consider episodic reinforcemen…
RL for Latent MDPs: Regret Guarantees and a Lower Bound
Jeongyeol Kwon, Yonathan Efroni, Constantine Caramanis +1
In this work, we consider the regret minimization problem for reinforcement learning in latent Markov Decision Processes (LMDP). In an LMDP, an MDP is randomly drawn from a set of…
On the computational and statistical complexity of over-parameterized matrix sensing
Jiacheng Zhuo, Jeongyeol Kwon, Nhat Ho +1
We consider solving the low rank matrix sensing problem with Factorized Gradient Descend (FGD) method when the true rank is unknown and over-specified, which we refer to as over-pa…
On the Minimax Optimality of the EM Algorithm for Learning Two-Component Mixed Linear Regression
Jeongyeol Kwon, Nhat Ho, Constantine Caramanis
We study the convergence rates of the EM algorithm for learning two-component mixed linear regression under all regimes of signal-to-noise ratio (SNR). We resolve a long-standing q…
The EM Algorithm gives Sample-Optimality for Learning Mixtures of Well-Separated Gaussians
Jeongyeol Kwon, Constantine Caramanis
We consider the problem of spherical Gaussian Mixture models with components when the components are well separated. A fundamental previous result established that separ…
EM Converges for a Mixture of Many Linear Regressions
Jeongyeol Kwon, Constantine Caramanis
We study the convergence of the Expectation-Maximization (EM) algorithm for mixtures of linear regressions with an arbitrary number of components. We show that as long as signa…
Global Convergence of EM Algorithm for Mixtures of Two Component Linear Regression
Jeongyeol Kwon, Wei Qian, Constantine Caramanis +2
The Expectation-Maximization algorithm is perhaps the most broadly used algorithm for inference of latent variable problems. A theoretical understanding of its performance, however…