Results 41 to 50 of about 788,437 (264)
, 2018 We present Residual Policy Learning (RPL): a simple method for improving nondifferentiable policies using model-free deep reinforcement learning. RPL thrives in complex robotic manipulation tasks where good but imperfect controllers are available. In these tasks, reinforcement learning from scratch remains data-inefficient or intractable, but learning ...
Silver, Tom +3 more
openaire +2 more sources
Residual Reinforcement Learning from Demonstrations
, 2021 Residual reinforcement learning (RL) has been proposed as a way to solve challenging robotic tasks by adapting control actions from a conventional feedback controller to maximize a reward signal. We extend the residual formulation to learn from visual inputs and sparse rewards using demonstrations.
Alakuijala, Minttu +4 more
openaire +2 more sources
Periodic residual learning for crowd flow forecasting
Proceedings of the 30th International Conference on Advances in Geographic Information Systems, 2022 Crowd flow forecasting, which aims to predict the crowds entering or leaving certain regions, is a fundamental task in smart cities. One of the key properties of crowd flow data is periodicity: a pattern that occurs at regular time intervals, such as a weekly pattern.
Wang, Chengxin, Liang, Yuxuan, Tan, Gary
openaire +2 more sources
Two-Stage Pansharpening Based on Multi-Level Detail Injection Network
IEEE Access, 2020 Pansharpening is an effective technology to obtain high resolution multispectral (HRMS) images by fusing low resolution multispectral (LRMS) images and high resolution panchromatic (PAN) images.
Jianwen Hu, Chenguang Du, Shaosheng Fan
doaj +1 more source
A Parallel Image Denoising Network Based on Nonparametric Attention and Multiscale Feature Fusion
SensorsConvolutional neural networks have achieved excellent results in image denoising; however, there are still some problems: (1) The majority of single-branch models cannot fully exploit the image features and often suffer from the loss of information.
Jing Mao +3 more
doaj +1 more source
SRflow: Deep learning based super-resolution of 4D-flow MRI data
Frontiers in Artificial Intelligence, 2022 Exploiting 4D-flow magnetic resonance imaging (MRI) data to quantify hemodynamics requires an adequate spatio-temporal vector field resolution at a low noise level. To address this challenge, we provide a learned solution to super-resolve in vivo 4D-flow
Suprosanna Shit +7 more
doaj +1 more source
Cascade Residual Learning: A Two-stage Convolutional Neural Network for Stereo Matching
, 2018 Leveraging on the recent developments in convolutional neural networks (CNNs), matching dense correspondence from a stereo pair has been cast as a learning problem, with performance exceeding traditional approaches.
Pang, Jiahao +4 more
core +1 more source
FR3 Path Loss in Outdoor Corridors: Physics-Guided Two-Ray Residual Learning
MathematicsAccurate path-loss characterization in the upper mid-band is critical for 5G/6G outdoor planning, yet classical deterministic expressions lose fidelity at 18 GHz, and purely data-driven regressors offer limited physical insight.
Jorge Celades-Martínez +4 more
doaj +1 more source
Gradient-Guided Residual Learning for Inverse Halftoning and Image Expanding
IEEE Access, 2020 Inverse halftoning and image expanding refer to problems to restore the pixel values of images from compressed images of smaller bit depth. Since these two problems are ill-posed, there are few perfect solutions.
Jin Yuan +5 more
doaj +1 more source
Low-Light Image Enhancement Using a Simple Network Structure
IEEE Access, 2023 Under low-light conditions, captured images can be affected by unsatisfactory lighting conditions. Low-light image enhancement called LLIE is a digital image processing to obtain natural normal-light images from the low-light image.
Takuro Matsui, Masaaki Ikehara
doaj +1 more source