Risk-sensitive Inverse Reinforcement Learning via Coherent Risk Models
By Anirudha Majumdar, Sumeet Singh, Ajay Mandlekar, Marco Pavone. 2017
72 citations recorded by the source corpus.
Subfields: Artificial Intelligence, General Decision Sciences, Management Science and Operations Research
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
- Sergey Levine
- Chelsea Finn
- Scott Niekum
- Pieter Abbeel
- Avi Singh
- Anca Dragan
- Daniel S. Brown
- Ken Goldberg
In General Decision Sciences
In Management Science and Operations Research
- Chelsea Finn
- Anca Dragan
- Sergey Levine
- Scott Niekum
- Avi Singh
- Larry Yang
- Kristian Hartikainen
- Ravi Srinivasan
Citations in the corpus
- End-to-End Robotic Reinforcement Learning without Reward Engineering (2019, background): “Data-driven approaches for reward specification [29, 1, 11, 17, 12, 48, 33, 9, 27, 4, 18] seek to overcome this issue, but typically require demonstration data to acquire rewards.”
- Bayesian Robust Optimization for Imitation Learning (2020, background): “[32] assume that risk-sensitive trajectories are available from a safe demonstrator and seek to optimize a policy that matches the risk-profile of this expert.”
- Planning and Decision-Making for Autonomous Vehicles (2018, background): “(121) devised a framework for risk-sensitive IRL to be able to take an expert’s risk sensitivity explicitly into account.”
- RAIL: Risk-Averse Imitation Learning (2017, methodology): “[Majumdar et al., 2017] take inspiration from studies like [Glimcher and Fehr, 2013, Shen et al.”
- Safe Imitation Learning via Fast Bayesian Reward Inference from Preferences (2020, background): “Other approaches to safe imitation learning seek to match the tail risk of the expert as well as find a policy that is indistinguishable from the demonstrations (Majumdar et al., 2017; Lacotte et al., 2019).”
- Policy Gradient Bayesian Robust Optimization for Imitation Learning (2021, background): “Risk-sensitive IRL methods (Lacotte et al., 2018; Majumdar et al., 2017; Santara et al., 2018) assume risk-averse experts and focus on optimizing policies that match the risk-aversion of the demonstrator; however, these methods focus on the aleatoric risk induced by transition probabilities and ther”