Nonamemanuscript No. (will be inserted by the editor) Reinforcement Learning to Adjust Parametrized Motor Primitives to

2011

Citations in the corpus

• End-to-End Driving Via Conditional Imitation Learning (2017, methodology): “Parameterized goals have been used to train motion controllers in robotics [7], [8], [18].”
• One-Shot Visual Imitation Learning via Meta-Learning (2017, background): “Another approach is to train controllers for a variety of tasks and learn a mapping from task representations to controller parameters [10, 25, 26].”
• Hindsight Experience Replay (2017, background): “Learning simultaneously policies for multiple tasks have been heavily explored in the context of policy search, e.g. Schmidhuber and Huber (1990); Caruana (1998); Da Silva et al. (2012); Kober et al. (2012); Devin et al. (2016); Pinto and Gupta (2016).”
• Temporal Difference Models: Model-Free Deep RL for Model-Based Control (2018, background): “Critically, unlike the works on contextual policies (Caruana, 1998; Da Silva et al., 2012; Kober et al., 2012) which require on-policy trajectories with each new goal, the value function approaches such as Horde (Sutton et al., 2011) and UVF (Schaul et al., 2015) can reuse off-policy data to learn…”
• Applying statistical generalization to determine search direction for reinforcement learning of movement primitives (2012, methodology): “[8] developed an algorithm where the control policy was improved by adapting a small set of global parameters, called meta-parameters.”
• MT-Opt: Continuous Multi-Task Robotic Reinforcement Learning at Scale (2021, methodology): “Combining multiple task policies has been explored in reinforcement learning by using gating networks [76, 50], conditioning policies on tasks [15], mapping tasks to parameters of a policy [13, 38, 82], distilling separate task policies into a shared multi-task policy [40, 77, 62, 53, 27, 5].”
• Neural networks and differential dynamic programming for reinforcement learning problems (2016, background): “Several attempts have been made to overcome this issue [11], [12].”
• Universal Value Function Approximators (2015, methodology): “…policies into a single situation-sensitive controller, for example parameterized skills (Da Silva et al., 2012) or parameterized motor primitives (Kober et al., 2012); and methods that directly parameterize a policy by the task as well as the state, e.g. using model-based RL (Deisenroth et al.,…”
• Learning to select and generalize striking movements in robot table tennis (2012, methodology): “A more detailed description and evaluation of this approach can be found in Kober et al. (2012).”
• Learning strategies in table tennis using inverse reinforcement learning (2014, background): “skills including even basic robot table tennis (Muelling et al. 2013; Kober et al. 2012).”
• Active choice of teachers, learning strategies and goals for a socially guided intrinsic motivation learner (2012, methodology): “Technically, we adopt a method of generalisation of policies for new outcomes similar to [15, 18].”
• Contextual Markov Decision Processes (2015, methodology): “A similar approach was used in learning meta-parameters of motor skills (Kober et al., 2012).”
• Towards transferring skills to flexible surgical robots with programming by demonstration and reinforcement learning (2016, methodology): “MODEL ONLINE REFINEMENT WITH REINFORCEMENT LEARNING In this section, an E-M based reinforcement learning algorithm called PoWER (Policy Learning by Weighting Exploration with the Returns) [24-27] is used to refine the IK model of the TSM manipulator to improve its tracking accuracy.”