Results 91 to 100 of about 7,347,743 (318)
Meta-Reinforcement Learning of Structured Exploration Strategies [PDF]
arXiv, 2018 Exploration is a fundamental challenge in reinforcement learning (RL). Many of the current exploration methods for deep RL use task-agnostic objectives, such as information gain or bonuses based on state visitation. However, many practical applications of RL involve learning more than a single task, and prior tasks can be used to inform how exploration
arxiv
QFS-space and its properties [PDF]
arXiv, 2022 In this paper, the concept of quasi-finitely separating map and quasiapproximate identity are introduced. Based on these concepts, QFS-spaces and quasicontinuous maps are defined. Properties and characterizations of QFS-spaces are explored. Main results are: (1) Each QFS-space is quasicontinuous space; (2) Closed subspaces, quasicontinuous projection ...
arxiv
Bootstrapping Parameter Space Exploration for Fast Tuning
International Conference on Supercomputing, 2018 The task of tuning parameters for optimizing performance or other metrics of interest such as energy, variability, etc. can be resource and time consuming. Presence of a large parameter space makes a comprehensive exploration infeasible.
Jayaraman J. Thiagarajan +9 more
semanticscholar +1 more source
Geometric Calibration of the ShadowCam Instrument on the Korea Pathfinder Lunar Orbiter
Journal of Astronomy and Space SciencesThe ShadowCam instrument on the Danuri spacecraft provides high-resolution views of shadowed portions of the Moon, which are illuminated by naturally scattered light from nearby sunlit terrain. The sensitive time-delay integration detector captures high
Emerson Jacob Speyerer +7 more
doaj +1 more source
Initial performance of the radio occultation experiment in the Venus orbiter mission Akatsuki
Earth, Planets and Space, 2017 After the arrival of Akatsuki spacecraft of Japan Aerospace Exploration Agency at Venus in December 2015, the radio occultation experiment, termed RS (Radio Science), obtained 19 vertical profiles of the Venusian atmosphere by April 2017.
Takeshi Imamura +38 more
doaj +1 more source
Cell-Free Latent Go-Explore [PDF]
arXiv, 2022 In this paper, we introduce Latent Go-Explore (LGE), a simple and general approach based on the Go-Explore paradigm for exploration in reinforcement learning (RL). Go-Explore was initially introduced with a strong domain knowledge constraint for partitioning the state space into cells.
arxiv
npj Microgravity, 2018 Future space exploration missions will take humans far beyond low Earth orbit and require complete crew autonomy. The ability to provide anaesthesia will be important given the expected risk of severe medical events requiring surgery.
M. Komorowski +3 more
semanticscholar +1 more source
Microscopic Phase-Space Exploration Modeling of ^{258}Fm Spontaneous Fission. [PDF]
Physical Review Letters, 2016 We show that the total kinetic energy (TKE) of nuclei after the spontaneous fission of ^{258}Fm can be well reproduced using simple assumptions on the quantum collective phase space explored by the nucleus after passing the fission barrier.
Y. Tanimura, D. Lacroix, S. Ayik
semanticscholar +1 more source
EXPLORING THE STACKING STATE-SPACE [PDF]
International Journal on Artificial Intelligence Tools, 2002 Nowadays, there is no doubt that machine learning techniques can be successfully applied to data mining tasks. Currently, the combination of several classifiers is one of the most active fields within inductive machine learning. Examples of such techniques are boosting, bagging and stacking.
Ledezma Espino, Agapito Ismael +2 more
openaire +3 more sources
Descent trajectory reconstruction and landing site positioning of Chang’E-4 on the lunar farside
Nature Communications, 2019 The Chang’E-4 mission in January 2019 had the major challenge to land on the lunar far side without traditional radiometric techniques due to the missing line-of-sight.
Jianjun Liu +16 more
doaj +1 more source