Amrit Singh Bedi

Direct Preference Optimization for Primitive-Enabled Hierarchical RL: A Bilevel Approach

Repair Aware Forgetting: An Iterative Approach to Unlearning in T2I Diffusion Models

TEST-TIME SCALING IN DIFFUSION LLMS VIA HIDDEN SEMI-AUTOREGRESSIVE EXPERTS

Multi-Level Multi-Turn RL Outperforms GRPO: Reasoning with Textual Feedback

Cut the Overcredit: Precision First Process Rewards for Reasoning LLMs

Advancing Regulation in Artificial Intelligence: An Auction-Based Approach

TRAM: Test-time Risk Adaptation with Mixture of Agents

Improved Sample Complexity Bounds For Diffusion Model Training Without Empirical Risk Minimizer Access

Mitigating Reward Hacking in Inference-Time Alignment of T2I Diffusion Models via Distributional Regularization

A Principled Approach to Chain-of-Thought Monitorability in Reasoning Models

SafeThink: A Key to Safety in Multi-Modal Large Reasoning Models

On the Global Optimality of Policy Gradient Methods in General Utility Reinforcement Learning

Does Thinking More Always Help? Mirage of Test-Time Scaling in Reasoning Models

On the Sample Complexity Bounds of Bilevel Reinforcement Learning

Bounded Rationality for LLMs: Satisficing Alignment at Inference-Time

SAIL: Self-improving Efficient Online Alignment of Large Language Models

Hierarchical Preference Optimization: Learning to achieve goals via feasible subgoals prediction

Towards Realistic Mechanisms That Incentivize Federated Participation and Contribution

Aligning Large Language Models With Preference Privacy

Auction-Based Regulation for Artificial Intelligence

LIAR: Leveraging Inverse Alignment to Jailbreak LLMs in Seconds

On the Sample Complexity of a Policy Gradient Algorithm with Occupancy Approximation for General Utility Reinforcement Learning

DIPPER: Direct Preference Optimization for Primitive-Enabled Hierarchical Reinforcement Learning

On the Global Convergence of RLHF Based Alignment With Neural Parametrization

AIME: AI System Optimization via Multiple LLM Evaluators