Inverse Reinforcement Learning with Learning and Leveraging Demonstrators' Varying Expertise Levels

S Oguchienti, M Ghasemi - 2023 59th Annual Allerton …, 2023 - ieeexplore.ieee.org
A common assumption in most Inverse Reinforcement Learning (IRL) methods is that human
demonstrations are drawn from an optimal policy. However, this assumption poses a …

[BOOK][B] Knowledge-Based Robot Sequential Decision-Making Under Partial Observability

S Amiri - 2022 - search.proquest.com
Robots frequently face complex tasks that require more than one action, where SDM
capabilities become necessary. SDM is challenging because actions and sensor …