31 citations · 58 across the 11 of their papers we have counts for
11 papers
Off-Context GRPO: Learning to Reason on Hard Problems using Privileged Information
Priyank Agrawal, Ankur Samanta, Shervin Ghasemlou +5
Reinforcement learning with verifiable rewards (RLVR) improves reasoning in large language models. Yet, typical RLVR approaches fail on difficult problems: when a model cannot gene…
Credit Assignment with Resets in Language Model Reasoning
Ankur Samanta, Akshayaa Magesh, Ayush Jain +7
Contemporary reinforcement learning with verifiable reward methods post-train language models on multi-step reasoning by assigning a single outcome reward uniformly across all toke…
Structure Enables Effective Self-Localization of Errors in LLMs
Ankur Samanta, Akshayaa Magesh, Ayush Jain +8
Self-correction in language models remains elusive. In this work, we explore whether language models can explicitly localize errors in incorrect reasoning, as a path toward buildin…
Iterative GRPO: Batch-Online Multi-Turn RL via Single-Turn RLHF
Daniel R. Jiang, Ankur Samanta, Yukai Yang +3
Practical LLM agents often operate over multi-turn conversations where success is determined only after the full interaction ends. Most multi-turn RL methods train via on-policy ro…
A Note on Code Quality Score: LLMs for Maintainable Large Codebases
Sherman Wong, Jalaj Bhandari, Leo Zhou Fan Yang +6
Maintaining code quality in large-scale software systems presents significant challenges, particularly in settings where a large numbers of engineers work concurrently on a codebas…
Aligned Multi Objective Optimization
Yonathan Efroni, Ben Kretzu, Daniel Jiang +4
To date, the multi-objective optimization literature has mainly focused on conflicting objectives, studying the Pareto front, or requiring users to balance tradeoffs. Yet, in machi…
Pearl: A Production-ready Reinforcement Learning Agent
Zheqing Zhu, Rodrigo de Salvo Braz, Jalaj Bhandari +12
Reinforcement learning (RL) is a versatile framework for optimizing long-term goals. Although many real-world problems can be formalized with RL, learning and deploying a performan…
Optimizing Long-term Value for Auction-Based Recommender Systems via On-Policy Reinforcement Learning
Ruiyang Xu, Jalaj Bhandari, Dmytro Korenkevych +4
Auction-based recommender systems are prevalent in online advertising platforms, but they are typically optimized to allocate recommendation slots based on immediate expected retur…
On Linear Convergence of Policy Gradient Methods for Finite MDPs
Jalaj Bhandari, Daniel Russo
We revisit the finite time analysis of policy gradient methods in the one of the simplest settings: finite state and action MDPs with a policy class consisting of all stochastic po…
Global Optimality Guarantees For Policy Gradient Methods
Jalaj Bhandari, Daniel Russo
Policy gradients methods apply to complex, poorly understood, control problems by performing stochastic gradient descent over a parameterized class of polices. Unfortunately, even…
A Finite Time Analysis of Temporal Difference Learning With Linear Function Approximation
Jalaj Bhandari, Daniel Russo, Raghav Singal
Temporal difference learning (TD) is a simple iterative algorithm used to estimate the value function corresponding to a given policy in a Markov decision process. Although TD is o…