10 papers
MAGIK: Mapping to Analogous Goals via Imagination-enabled Knowledge Transfer
Ajsal Shereef Palattuparambil, Thommen George Karimpanal, Santu Rana
Humans excel at analogical reasoning - applying knowledge from one task to a related one with minimal relearning. In contrast, reinforcement learning (RL) agents typically require…
Shared Latent Structures Enable Unified Backdoor Detection and Mitigation in LLMs
Omar Mahmoud, Aly M. Kassem, Thommen George Karimpanal +4
Backdoor attacks in large language models (LLMs) are often treated as isolated trigger-response failures, motivating defenses tailored to specific triggers or behaviors. We show th…
Leveraging Human Feedback for Semantically-Relevant Skill Discovery
Maxence Hussonnois, Thommen George Karimpanal, Santu Rana
Unsupervised skill discovery in reinforcement learning aims to intrinsically motivate agents to discover diverse and useful behaviours. However, unconstrained approaches can produc…
ASPECT:Analogical Semantic Policy Execution via Language Conditioned Transfer
Ajsal Shereef Palattuparambil, Thommen George Karimpanal, Santu Rana
Reinforcement Learning (RL) agents often struggle to generalize knowledge to new tasks, even those structurally similar to ones they have mastered. Although recent approaches have…
The Unintended Trade-off of AI Alignment:Balancing Hallucination Mitigation and Safety in LLMs
Omar Mahmoud, Ali Khalil, Buddhika Laknath Semage +2
Hallucination in large language models (LLMs) has been widely studied in recent years, with progress in both detection and mitigation aimed at improving truthfulness. Yet, a critic…
Dynamic Policy Fusion for User Alignment Without Re-Interaction
Ajsal Shereef Palattuparambil, Thommen George Karimpanal, Santu Rana
Deep reinforcement learning (RL) policies, although optimal in terms of task rewards, may not align with the personal preferences of human users. To ensure this alignment, a naive…