Efficient model-based reinforcement learning through optimistic policy search and planning
Abstract Model-based reinforcement learning algorithms with probabilistic dynamical
models are amongst the most data-efficient learning methods. This is often attributed to their …
models are amongst the most data-efficient learning methods. This is often attributed to their …
Where to go next: Learning a subgoal recommendation policy for navigation in dynamic environments
Robotic navigation in environments shared with other robots or humans remains
challenging because the intentions of the surrounding agents are not directly observable …
challenging because the intentions of the surrounding agents are not directly observable …
Model-augmented actor-critic: Backpropagating through paths
Current model-based reinforcement learning approaches use the model simply as a learned
black-box simulator to augment the data for policy optimization or value function learning. In …
black-box simulator to augment the data for policy optimization or value function learning. In …
Model-based reinforcement learning for semi-markov decision processes with neural odes
J Du, J Futoma, F Doshi-Velez - Advances in Neural …, 2020 - proceedings.neurips.cc
We present two elegant solutions for modeling continuous-time dynamics, in a novel model-
based reinforcement learning (RL) framework for semi-Markov decision processes (SMDPs) …
based reinforcement learning (RL) framework for semi-Markov decision processes (SMDPs) …
[HTML][HTML] Machine learning meets advanced robotic manipulation
Automated industries lead to high quality production, lower manufacturing cost and better
utilization of human resources. Robotic manipulator arms have major role in the automation …
utilization of human resources. Robotic manipulator arms have major role in the automation …
Learning interaction-aware guidance for trajectory optimization in dense traffic scenarios
B Brito, A Agarwal… - IEEE Transactions on …, 2022 - ieeexplore.ieee.org
Autonomous navigation in dense traffic scenarios remains challenging for autonomous
vehicles (AVs) because the intentions of other drivers are not directly observable and AVs …
vehicles (AVs) because the intentions of other drivers are not directly observable and AVs …
The role of lookahead and approximate policy evaluation in reinforcement learning with linear value function approximation
A Winnicki, J Lubars, M Livesay… - Operations …, 2024 - pubsonline.informs.org
Function approximation is widely used in reinforcement learning to handle the
computational difficulties associated with very large state spaces. However, function …
computational difficulties associated with very large state spaces. However, function …
PAC-NMPC with Learned Perception-Informed Value Function
A Polevoy, M Gonzales, M Kobilarov… - arXiv preprint arXiv …, 2023 - arxiv.org
Nonlinear model predictive control (NMPC) is typically restricted to short, finite horizons to
limit the computational burden of online optimization. This makes a global planner …
limit the computational burden of online optimization. This makes a global planner …
Epistemic Uncertainty for Practical Deep Model-Based Reinforcement Learning
S Curi - 2022 - research-collection.ethz.ch
Reinforcement Learning (RL) has advanced the state-of-the-art in many applications in the
last decade. The root of its success stems from having access to high-quality simulators …
last decade. The root of its success stems from having access to high-quality simulators …
[PDF][PDF] Learning a Guidance Policy from Humans for Social Navigation
Autonomous mobile robots navigating among humans must not only consider safety and
efficiency but also move acceptably in the current social context. A hybrid deep …
efficiency but also move acceptably in the current social context. A hybrid deep …