Reinforcement

By Ianis Lallemand, Andrew Y. Ng, Adam Coates, Octobre, Mark Diel, Varun Ganapathi, Jamie Schulte, B. Tse, Eric Berger, Eric Liang. 2018

Subfields: Reinforcement Learning, Robotics & Control

Citations in the corpus

• Deep Imitation Learning for Complex Manipulation Tasks from Virtual Reality Teleoperation (2017, neutral): “, [47], [48], [49]) work in state space, more recent work has been able to learn deep neural net policies from pixels to actions (e.”
• Hindsight Experience Replay (2017, neutral): “…et al., 2015), and defeating the best human player at the game of Go (Silver et al., 2016), as well as robotic tasks such as helicopter control (Ng et al., 2006), hitting a baseball (Peters and Schaal, 2008), screwing a cap onto a bottle (Levine et al., 2015), or door opening (Chebotar et al.,…”
• Evolution Strategies as a Scalable Alternative to Reinforcement Learning (2017, neutral): “Successes of this approach include systems that learn to play Atari from pixels [Mnih et al., 2015], perform helicopter aerobatics Ng et al. [2006], or play expert-level Go [Silver et al., 2016].”
• Ray: A Distributed Framework for Emerging AI Applications (2017, neutral): “RL-based systems have already yielded remarkable results, such as Google’s AlphaGo beating a human world champion [54], and are beginning to find their way into dialogue systems, UAVs [42], and robotic manipulation [25, 60].”
• A Brief Survey of Deep Reinforcement Learning (2017, neutral): “Although RL had some successes in the past [119, 109, 51, 79], previous approaches lacked scalablity and were inherently limited to fairly low-dimensional problems.”
• Active Reward Learning (2014, neutral): “It does not rely on exhaustive sampling of the state-action space and, as a result, many recent advances in robot learning rely on policy search [22, 20, 26].”
• A robot exploration strategy based on Q-learning network (2016, unknown): “RL was also applied on an autonomous helicopter flight [8] and autonomous inverted helicopter flight [9], by collecting the flight data and learning a non-linear model of the aerodynamics.”
• Black-box data-efficient policy search for robotics (2017, unknown): “Early examples of such model-based PS include applications on helicopter hovering [25], [26] and blimp control [28].”
• Model learning for robot control: a survey (2011, unknown): “Forward models have also been essential in model based reinforcement learning approaches, which relate to the problem of optimal control [156,9, 103].”
• What-If Reasoning with Counterfactual Gaussian Processes (2017, unknown): “In model-based RL, a model of an action’s effect on the subsequent state is produced as a by-product either offline before optimizing the policy (e.g., Ng et al. 2006) or incrementally as the agent interacts with its environment.”
• Imitation Learning (2017, unknown): “These include flying vehicles [Sammut et al. 2014; Abbeel et al. 2007; Ng et al. 2006] and ground vehicles [Silver et al. 2008; Ratliff et al. 2007; Chernova and Veloso 2007a; Ollis et al. 2007].”
• Efficient reinforcement learning for robots using informative simulated priors (2015, unknown): “While RL has shown significant promise and impressive results in some domains, such as robot locomotion [2], helicopter flight [3], [4], backgammon [5], and elevator control [6], RL has not yet, in general, been employed in mainstream robotics as a standard tool for controller design.”
• Reinforcement learning for robot soccer (2009, unknown): “2000; Peters and Schaal 2008b) has led to the development of several impressive real-world demonstrators including autonomously flying artistic maneuvers with a helicopter (Bagnell and Schneider 2001; Ng et al. 2004), teaching a robot to hit a baseball (Peters and Schaal 2006), operational space con”