Results 91 to 100 of about 5,876,040 (331)

Using Inverse Reinforcement Learning with Real Trajectories to Get More Trustworthy Pedestrian Simulations

Mathematics, 2020
Reinforcement learning is one of the most promising machine learning techniques to get intelligent behaviors for embodied agents in simulations. The output of the classic Temporal Difference family of Reinforcement Learning algorithms adopts the form of ...
Francisco Martinez-Gil, Miguel Lozano, Ignacio García-Fernández, Pau Romero, Dolors Serra, Rafael Sebastián   +5 more
doaj   +1 more source

Reinforcement learning-based link adaptation in long delayed underwater acoustic channel [PDF]

MATEC Web of Conferences, 2019
In this paper, we apply reinforcement learning, a significant area of machine learning, to formulate an optimal self-learning strategy to interact in an unknown and dynamically variable underwater channel.
Wang Jingxi, Yuen Chau, Guan Yong Liang, Ge Fengxiang   +3 more
doaj   +1 more source

Stochastic Inverse Reinforcement Learning

, 2020
The goal of the inverse reinforcement learning (IRL) problem is to recover the reward functions from expert demonstrations. However, the IRL problem like any ill-posed inverse problem suffers the congenital defect that the policy may be optimal for many ...
Ju, Ce
core  

Nanoparticle‐Coated X2CrNiMo17‐12‐2 Powder for Additive Manufacturing – Part I: Surface, Flowability, and Optical Properties of SiC, Si, and Si3N4 Coated Metal Powders

Advanced Engineering Materials, EarlyView.
Herein, silicon‐based nanoparticle coatings on X2CrNiMo17‐12‐2 metal powder are presented. The coating process scale, process parameters, nanoparticle size (65–200 nm) as well as the coating amount are discussed regarding powder properties. The surface roughness affects the flowability, while reflectance depends on the coating material and surface ...
Arne Lüddecke, Nick Hantke, Harald Zetzener, Jan T. Sehrt, Arno Kwade   +4 more
wiley   +1 more source

Optimizing Reinforcement Learning Using a Generative Action-Translator Transformer

Algorithms
In recent years, with the rapid advancements in Natural Language Processing (NLP) technologies, large models have become widespread. Traditional reinforcement learning algorithms have also started experimenting with language models to optimize training ...
Jiaming Li, Ning Xie, Tingting Zhao
doaj   +1 more source

Beyond Order: Perspectives on Leveraging Machine Learning for Disordered Materials

Advanced Engineering Materials, EarlyView.
This article explores how machine learning (ML) revolutionizes the study and design of disordered materials by uncovering hidden patterns, predicting properties, and optimizing multiscale structures. It highlights key advancements, including generative models, graph neural networks, and hybrid ML‐physics methods, addressing challenges like data ...
Hamidreza Yazdani Sarvestani, Surabhi Nadigotti, Erfan Fatehi, Derek Aranguren van Egmond, Behnam Ashrafi   +4 more
wiley   +1 more source

Reinforcement Learning

, 2019
Reinforcement Learning (RL) is one of the model free machine learning algorithms where the agent learns its behaviours from the environment by actually interacting with it. This is better than the offline planner because the agent actually interacts with the environment to learn its behaviours because it is almost impossible to simulate a real world in
Jimut Bahan Pal, Debadri Chatterjee, Sounak Modak   +2 more
openaire   +2 more sources

Superposition-Inspired Reinforcement Learning and Quantum Reinforcement Learning [PDF]

, 2008
According to the existing problems in RL area, such as low learning speed and tradeoff between exploration and exploitation, SIRL and QRL methods are introduced based on the theory of RL and quantum computation in this chapter, which follows the developing roadmap from the superposition-inspired methods to the RL methods in quantum systems.
Dao-Yi Dong, Chun-Lin Chen
openaire   +3 more sources

Universal Reinforcement Learning [PDF]

IEEE Transactions on Information Theory, 2010
We consider an agent interacting with an unmodeled environment. At each time, the agent makes an observation, takes an action, and incurs a cost. Its actions can influence future observations and costs. The goal is to minimize the long-term average cost.
Farias, Vivek F., Moallemi, Ciamac C., Van Roy, Benjamin, Weissman, Tsachy   +3 more
openaire   +4 more sources

Statistical Distributions of Morphologically Classified Defects in Metal Additive Manufacturing with Implications for Fatigue Life Prediction

Advanced Engineering Materials, EarlyView.
Morphological features of three defect types in metal additive manufacturing (AM)—lack of fusion, keyhole, and gas‐entrapped pores—are statistically characterized using best‐fit distributions evaluated via coefficient‐of‐determination, Kolmogorov–Smirnov test, and quantile–quantile plots.
Ahmad Serjouei, Golnaz Shahtahmassebi
wiley   +1 more source