Dynamic multi-objective optimisation using deep reinforcement learning: benchmark, algorithm and an application to identify vulnerable zones based on water quality
Dynamic multi-objective optimisation problem (DMOP) has brought a great challenge to the
reinforcement learning (RL) research area due to its dynamic nature such as objective …
reinforcement learning (RL) research area due to its dynamic nature such as objective …
Robust Black-Box Optimization for Stochastic Search and Episodic Reinforcement Learning
M Hüttenrauch, G Neumann - Journal of Machine Learning Research, 2024 - jmlr.org
Black-box optimization is a versatile approach to solve complex problems where the
objective function is not explicitly known and no higher order information is available. Due to …
objective function is not explicitly known and no higher order information is available. Due to …
Geometric reinforcement learning for robotic manipulation
Reinforcement learning (RL) is a popular technique that allows an agent to learn by trial and
error while interacting with a dynamic environment. The traditional Reinforcement Learning …
error while interacting with a dynamic environment. The traditional Reinforcement Learning …
A probabilistic model-based online learning optimal control algorithm for soft pneumatic actuators
Soft robots are increasingly being employed in different fields and various designs are
created to satisfy relevant requirements. The wide ranges of design bring challenges to soft …
created to satisfy relevant requirements. The wide ranges of design bring challenges to soft …
Safeapt: Safe simulation-to-real robot learning using diverse policies learned in simulation
The framework of sim-to-real learning, ie, training policies in simulation and transferring
them to real-world systems, is one of the most promising approaches towards data-efficient …
them to real-world systems, is one of the most promising approaches towards data-efficient …
Empirical prior based probabilistic inference neural network for policy learning
Y Li, S Guo, Z Gan - Information Sciences, 2022 - Elsevier
Reinforcement learning is very much democratized for autonomous control of an unknown
dynamics system. However, low data efficiency is a practical concern in physical systems …
dynamics system. However, low data efficiency is a practical concern in physical systems …
Data-efficient reinforcement learning for variable impedance control
One of the most crucial steps toward achieving human-like manipulation skills in robots is to
incorporate compliance into the robot controller. Compliance not only makes the robot's …
incorporate compliance into the robot controller. Compliance not only makes the robot's …
Low-cost wireless modular soft tensegrity robots
Completely soft robots are emerging as a compelling new platform for exploring and
operating in unstructured, rugged, and dynamic environments. Unfortunately, the very …
operating in unstructured, rugged, and dynamic environments. Unfortunately, the very …
Data-efficient model learning and prediction for contact-rich manipulation tasks
In this letter, we investigate learning forward dynamics models and multi-step prediction of
state variables (long-term prediction) for contact-rich manipulation. The problems are …
state variables (long-term prediction) for contact-rich manipulation. The problems are …
Learning virtual grasp with failed demonstrations via bayesian inverse reinforcement learning
We propose Bayesian Inverse Reinforcement Learning with Failure (BIRLF), which makes
use of failed demonstrations that were often ignored or filtered in previous methods due to …
use of failed demonstrations that were often ignored or filtered in previous methods due to …