Influencing towards stable multi-agent interactions

WZ Wang, A Shih, A **e… - Conference on robot …, 2022 - proceedings.mlr.press
Learning in multi-agent environments is difficult due to the non-stationarity introduced by an
opponent's or partner's changing behaviors. Instead of reactively adapting to the other …

Blind spot detection for safe sim-to-real transfer

R Ramakrishnan, E Kamar, D Dey, E Horvitz… - Journal of Artificial …, 2020 - jair.org
Agents trained in simulation may make errors when performing actions in the real world due
to mismatches between training and execution environments. These mistakes can be …

[HTML][HTML] Deep Reinforcement Learning for sim-to-real policy transfer of VTOL-UAVs offshore docking operations

AM Ali, A Gupta, HA Hashim - Applied Soft Computing, 2024 - Elsevier
This paper proposes a novel Reinforcement Learning (RL) approach for sim-to-real policy
transfer of Vertical Take-Off and Landing Unmanned Aerial Vehicle (VTOL-UAV). The …

[PDF][PDF] Zero-shot skill composition and simulation-to-real transfer by learning task representations

Z He, R Julian, E Heiden, H Zhang… - arxiv preprint arxiv …, 2018 - eric-heiden.com
Simulation-to-real transfer is an important strategy for making reinforcement learning
practical with real robots. Successful sim-to-real transfer systems have difficulty producing …

Conditionally Combining Robot Skills using Large Language Models

KR Zentner, R Julian, B Ichter… - 2024 IEEE International …, 2024 - ieeexplore.ieee.org
This paper combines two contributions. First, we introduce an extension of the Meta-World
benchmark, which we call" Language-World," which allows a large language model to …

Efficient multi-task learning via iterated single-task transfer

KR Zentner, U Puri, Y Zhang, R Julian… - 2022 IEEE/RSJ …, 2022 - ieeexplore.ieee.org
In order to be effective general purpose machines in real world environments, robots not
only will need to adapt their existing manipulation skills to new circumstances, they will need …

Hierarchical End-to-end Control Policy for Multi-degree-of-freedom Manipulators

CH Min, JB Song - International Journal of Control, Automation and …, 2022 - Springer
In recent years, several control policies for a multi-degree-of-freedom (DOF) manipulator
using deep reinforcement learning have been proposed. To avoid complexity, previous …

A simple approach to continual learning by transferring skill parameters

KR Zentner, R Julian, U Puri, Y Zhang… - arxiv preprint arxiv …, 2021 - arxiv.org
In order to be effective general purpose machines in real world environments, robots not
only will need to adapt their existing manipulation skills to new circumstances, they will need …

Scaling simulation-to-real transfer by learning a latent space of robot skills

RC Julian, E Heiden, Z He, H Zhang… - … Journal of Robotics …, 2020 - journals.sagepub.com
We present a strategy for simulation-to-real transfer, which builds on recent advances in
robot skill decomposition. Rather than focusing on minimizing the simulation–reality gap, we …

Auto-conditioned recurrent mixture density networks for learning generalizable robot skills

H Zhang, E Heiden, S Nikolaidis, JJ Lim… - arxiv preprint arxiv …, 2018 - arxiv.org
Personal robots assisting humans must perform complex manipulation tasks that are
typically difficult to specify in traditional motion planning pipelines, where multiple objectives …