Generalized model learning for Reinforcement Learning on a humanoid robot

By Todd Hester, Michael J. Quinlan, Peter Stone. 2010

113 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Biomedical Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Biomedical Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Conservative Q-Improvement: Reinforcement Learning for an Interpretable Decision-Tree Policy (2019, methodology): “on (as in our and other methods) and temporal abstraction. [Pyeatt et al., 2001] is a work which uses a tree as an internal structure and incrementally builds the tree, splitting on Q-values. Unlike [Hester et al., 2010], an approach that uses decision trees as part of the policy representation, we ”
Reinforcement learning in robotics: A survey (2013, background): “to learn a penalty kick on a humanoid robot (Hester et al., 2010).”
Towards High Confidence Off-Policy Reinforcement Learning for Clinical Applications (2018, background): “They are more sample efficient (Hester et al., 2010; Hallak et al., 2015), however it may be difficult to quantify the approximation error or the model bias of these methods.”
Transfer learning for reinforcement learning on a physical robot (2010, methodology): “A limited amount of previous work has been done using the Nao, but this work focused on simulation work, with only a single run on a physical robot [7].”
Value and reward based learning in neurorobots (2013, methodology): “Although many model-based methods includes model learning (Deisenroth et al., 2009; Hester et al., 2010) have been proposed in this field, they compute an optimal state value function which is a solution of a non-linear Bellman’s equation.”
Learning and Using Models (2012, methodology): “This algorithm has been shown to be effective on gridworld tasks such as Taxi, as well as on humanoid robots learning to score penalty kicks (Hester et al, 2010).”
The Open-Source TEXPLORE Code Release for Reinforcement Learning on Robots (2013, methodology): “In addition to learning to control the velocity of an autonomous vehicle, a variant of texplore has also been used to learn how to score penalty kicks in the Standard Platform League of RoboCup [6].”
Humanoid robots learning to walk faster: from the real world to simulation and back (2013, background): “More recently, there has been progress in learning on bipedal robots, including learning decision tree models to score penalty kicks [11], learning to walk by demonstration [18], and improving walk stability by learning from human feedback [19].”
Learning Powerful Kicks on the Aibo ERS-7: The Quest for a Striker (2010, methodology): “[8] recently applied the RL-DT model-based reinforcement learning algorithm to the task of scoring penalty goals.”