15 citations · 37 across the 28 of their papers we have counts for
20 papers · 1 filter
FlowBalance: Verifier-Grounded Self-Improvement from On-Policy Reasoning Experience
Zixun Huang, Kishan Panaganti, Haitao Mi +1
A reasoning model can improve from its own on-policy experience, but this inner loop is fragile: terminal verifiers provide reliable yet sparse supervision, while dense same-model…
Stale but Stable: Staleness-Adaptive Trust Regions for Stabilizing Asynchronous Reinforcement Learning
Junyao Yang, Yucheng Shi, Zongxia Li +6
Asynchronous reinforcement learning improves throughput by decoupling rollout generation from optimization, but the resulting staleness is an inevitable byproduct, compounded joint…
Group Distributionally Robust Optimization-Driven Reinforcement Learning for LLM Reasoning
Kishan Panaganti, Zhenwen Liang, Wenhao Yu +2
Recent progress in Large Language Model (LLM) reasoning is increasingly driven by the refinement of post-training loss functions and alignment strategies. However, standard Reinfor…
Can LLMs Guide Their Own Exploration? Gradient-Guided Reinforcement Learning for LLM Reasoning
Zhenwen Liang, Sidi Lu, Wenhao Yu +4
Reinforcement learning has become essential for strengthening the reasoning abilities of large language models, yet current exploration mechanisms remain fundamentally misaligned w…
Every Question Has Its Own Value: Reinforcement Learning with Explicit Human Values
Dian Yu, Yulai Zhao, Kishan Panaganti +3
We propose Reinforcement Learning with Explicit Human Values (RLEV), a method that aligns Large Language Model (LLM) optimization directly with quantifiable human value signals. Wh…
Evolving Language Models without Labels: Majority Drives Selection, Novelty Promotes Variation
Yujun Zhou, Zhenwen Liang, Haolin Liu +7
Large language models (LLMs) can improve through reinforcement learning with verifiable rewards (RLVR), but many settings lack reference answers or external verifiers. Label-free s…
Risk-Averse Total-Reward Reinforcement Learning
Xihong Su, Jia Lin Hau, Gersi Doko +2
Risk-averse total-reward Markov Decision Processes (MDPs) offer a promising framework for modeling and solving undiscounted infinite-horizon objectives. Existing model-based algori…
Efficient Policy Optimization in Robust Constrained MDPs with Iteration Complexity Guarantees
Sourav Ganguly, Kishan Panaganti, Arnob Ghosh +1
Constrained decision-making is essential for designing safe policies in real-world control systems, yet simulated environments often fail to capture real-world adversities. We cons…
KL-regularization Itself is Differentially Private in Bandits and RLHF
Yizhou Zhang, Kishan Panaganti, Laixi Shi +2
Differential Privacy (DP) provides a rigorous framework for privacy, ensuring the outputs of data-driven algorithms remain statistically indistinguishable across datasets that diff…
Robust LLM Alignment via Distributionally Robust Direct Preference Optimization
Zaiyan Xu, Sushil Vemuri, Kishan Panaganti +3
A major challenge in aligning large language models (LLMs) with human preferences is the issue of distribution shift. LLM alignment algorithms rely on static preference datasets, a…
Hybrid Transfer Reinforcement Learning: Provable Sample Efficiency from Shifted-Dynamics Data
Chengrui Qu, Laixi Shi, Kishan Panaganti +2
Online Reinforcement learning (RL) typically requires high-stakes online interaction data to learn a policy for a target task. This prompts interest in leveraging historical data t…
Distributionally Robust Constrained Reinforcement Learning under Strong Duality
Zhengfei Zhang, Kishan Panaganti, Laixi Shi +3
We study the problem of Distributionally Robust Constrained RL (DRC-RL), where the goal is to maximize the expected reward subject to environmental distribution shifts and constrai…
Model-Free Robust -Divergence Reinforcement Learning Using Both Offline and Online Data
Kishan Panaganti, Adam Wierman, Eric Mazumdar
The robust -regularized Markov Decision Process (RRMDP) framework focuses on designing control policies that are robust against parameter uncertainties due to mismatches between…
Bridging Distributionally Robust Learning and Offline RL: An Approach to Mitigate Distribution Shift and Partial Data Coverage
Kishan Panaganti, Zaiyan Xu, Dileep Kalathil +1
The goal of an offline reinforcement learning (RL) algorithm is to learn optimal polices using historical (offline) data, without access to the environment for online exploration.…
Improved Sample Complexity Bounds for Distributionally Robust Reinforcement Learning
Zaiyan Xu, Kishan Panaganti, Dileep Kalathil
We consider the problem of learning a control policy that is robust against the parameter mismatches between the training environment and testing environment. We formulate this as…
Personalized Reward Learning with Interaction-Grounded Learning (IGL)
Jessica Maghakian, Paul Mineiro, Kishan Panaganti +3
In an era of countless content offerings, recommender systems alleviate information overload by providing users with personalized content suggestions. Due to the scarcity of explic…
Robust Reinforcement Learning using Offline Data
Kishan Panaganti, Zaiyan Xu, Dileep Kalathil +1
The goal of robust reinforcement learning (RL) is to learn a policy that is robust against the uncertainty in model parameters. Parameter uncertainty commonly occurs in many real-w…
Sample Complexity of Robust Reinforcement Learning with a Generative Model
Kishan Panaganti, Dileep Kalathil
The Robust Markov Decision Process (RMDP) framework focuses on designing control policies that are robust against the parameter uncertainties due to the mismatches between the simu…
Robust Reinforcement Learning using Least Squares Policy Iteration with Provable Performance Guarantees
Kishan Panaganti, Dileep Kalathil
This paper addresses the problem of model-free reinforcement learning for Robust Markov Decision Process (RMDP) with large state spaces. The goal of the RMDP framework is to find a…
Bounded Regret for Finitely Parameterized Multi-Armed Bandits
Kishan Panaganti, Dileep Kalathil
We consider the problem of finitely parameterized multi-armed bandits where the model of the underlying stochastic environment can be characterized based on a common unknown parame…