RDIRL is an online deep inverse reinforcement learning method that updates a learned cost after each expert demonstration with a Kalman-style second-order Newton step, and it outperforms batch IRL baselines in simulated control and radar tasks.
The goal is to iteratively updateθ such that trajectories generated from the current policyq(τ) match the expert demonstrations
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Recursive Deep Inverse Reinforcement Learning
RDIRL is an online deep inverse reinforcement learning method that updates a learned cost after each expert demonstration with a Kalman-style second-order Newton step, and it outperforms batch IRL baselines in simulated control and radar tasks.