MOCI jointly infers shared constraints and individual preferences from heterogeneous expert trajectories via multi-objective inverse reinforcement learning and outperforms baselines on grid-world predictive performance.
Learning shared safety constraints from multi-task demonstrations.Advances in Neural Information Processing Systems, 36:5808–5826
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 2roles
background 1polarities
background 1representative citing papers
Agentic safety fails to generalize across tasks because the task-to-safe-controller mapping has a higher Lipschitz constant than the task-to-controller mapping alone, as proven in linear-quadratic control and demonstrated in quadcopter and LLM experiments.
citing papers explorer
-
Multi-Objective Constraint Inference using Inverse reinforcement learning
MOCI jointly infers shared constraints and individual preferences from heterogeneous expert trajectories via multi-objective inverse reinforcement learning and outperforms baselines on grid-world predictive performance.
-
Why Does Agentic Safety Fail to Generalize Across Tasks?
Agentic safety fails to generalize across tasks because the task-to-safe-controller mapping has a higher Lipschitz constant than the task-to-controller mapping alone, as proven in linear-quadratic control and demonstrated in quadcopter and LLM experiments.