Scaling Up Approximate Value Iteration with Options: Better Policies with Fewer Iterations
By Timothy A. Mann, Shie Mannor. 2014
59 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
- Sergey Levine
- Doina Precup
- Shie Mannor
- Daniel Jaymin Mankowitz
- Pieter Abbeel
- Khimya Khetarpal
- David Abel
- Martin Klissarov
Citations in the corpus
- Why Does Hierarchy (Sometimes) Work So Well in Reinforcement Learning? (2019, unknown): “Previous work has also analyzed HRL (specifically, the options framework) in the more general setting of continuous states (Mann & Mannor, 2014).”
- Time-regularized interrupting options (2014, background): “(b) Option interruption using regularization which preserves option duration and results in a more direct and efficient planning solution.”
- Iterative Hierarchical Optimization for Misspecified Problems (IHOMP) (2016, background): “Examples include inventory management systems [Mann and Mannor, 2014] as well as maintenance scheduling of generation units and transmission lines in smart grids [Abiri-Jahromi et al., 2013, Wu et al., 2010].”
- Adaptive Skills Adaptive Partitions (ASAP) (2016, background): “The solution to a MDP is a policy π : X → ∆A which is a function mapping states to a probability distribution over actions.”
- Value Preserving State-Action Abstractions (2020, background): “, 2019), and make planning more e cient (Mann and Mannor, 2014; Mann et al., 2015).”
- Options of Interest: Temporal Abstraction with Interest Functions (2020, background): “Options in particular have been shown to speed up convergence both empirically (Precup 2000) and theoretically (Mann and Mannor 2014).”
- Learning Robust Options (2018, background): “…(da Silva, Konidaris, and Barto 2012; Mankowitz, Mann, and Mannor 2016b; 2016a) or macro-actions (Hauskrecht et al. 1998) have been shown both theoretically (Precup, Sutton, and Singh 1998) and experimentally (Mann and Mannor 2014) to result in faster convergence rates in RL planning algorithms.”
- DREAM Architecture: a Developmental Approach to Open-Ended Learning in Robotics (2020, unknown): “It can accelerate learning [107] or make learning more robust [108].”