Policy gradients in linearly-solvable MDPs

By E. Todorov. 2010

18 citations recorded by the source corpus.

Subfields: Reinforcement Learning

How researchers see this paper

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

In Reinforcement Learning

1. Sergey Levine
2. Ofir Nachum
3. Mohammad Norouzi
4. Dale Schuurmans
5. Pieter Abbeel
6. George Tucker
7. Chelsea Finn
8. Shixiang Shane Gu

Citations in the corpus

• Reinforcement Learning and Control as Probabilistic Inference: Tutorial and Review (2018, methodology): “The motivation for linearly solvable MDPs is typically based on computationally tractable exact solutions for tabular settings (Todorov, 2010), which are enabled essentially by dispensing with the non-linear maximization operator in the standard RL framework.”
• Bridging the Gap Between Value and Policy Based Reinforcement Learning (2017, methodology): “Even though entropy regularized policy optimization is a well studied topic in RL [46, 39, 40, 47, 5, 4, 6, 7]–in fact, one that has been attracting renewed interest from concurrent work [25, 11]–we contribute new observations to this study that are essential for the methods we propose: first, we id”
• Variational Policy Search via Trajectory Optimization (2013, unknown): “However, we can also obtain a Gaussian q(ζ) by using the framework of linearly solvable MDPs [16] and the closely related concept of maximum entropy control [23].”