Conquering coupled diabatic potential energy surfaces with artificial neural networks for nonadiabatic dynamics

Autor: Williams, David M. G.

The Born-Oppenheimer (BO) approximation is a cornerstone of the theoretical treatment of molecular processes. It separates nuclear and electronic motion, splitting up an immensely challenging many-body problem into two highly involved but well-unders

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=od_______166::31738a6b59db5f67d0dc79b26f827c30
https://hal.science/tel-03917774

Machine Learning for Human Action Recognition and Pose Estimation based on 3D Information

Autor: Luvizon, Diogo

Publikováno v: Computer Vision and Pattern Recognition [cs.CV]. Cergy Paris Université, 2019. English

3D human action recognition is a challenging task due to the complexity of human movementsand to the variety on poses and actions performed by distinct subjects. Recent technologies basedon depth sensors can provide 3D human skeletons with low comput

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=od_______212::2211b7d905bd3e3e823fb4000883bf7b
https://tel.archives-ouvertes.fr/tel-02492463/document

LSTM Path-Maker: a new LSTM-based strategy for the multi-agent patrolling

Autor: Mehdi Othmani-Guibourg, Jean-Loup Farges, Amal El Fallah Seghrouchni

Publikováno v: HICSS
Proceedings of the 52nd Hawaii International Conference on System Sciences
52nd Annual Hawaii International Conference on System Sciences (HICSS 2019)
52nd Annual Hawaii International Conference on System Sciences (HICSS 2019), Jan 2019, Wailea, HI, United States. pp.616-625

International audience; For over a decade, the multi-agent patrolling task has received a growing attention from the multi-agent community due to its wide range of potential applications. Various algorithms based on reactive and cognitive architectur

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::70305d18300123d4e5278a825994b239
https://doi.org/10.24251/hicss.2019.076

Génération de chemin à partir de réseaux de neurones récurrents LSTM dans le contexte de la Patrouille Multi-agent

Autor: Amal El Fallah-Seghrouchni, Jean-Loup Farges, Mehdi Othmani-Guibourg

Publikováno v: 30th International Conference on Tools with Artificial Intelligence (ICTAI)
30th International Conference on Tools with Artificial Intelligence (ICTAI), Nov 2018, VOLOS, Greece
ICTAI

International audience; We propose a conceptually simple new decentralised and non-communicating strategy for the multi-agent patrolling based on the LSTM architecture. The recurrent neural networks and more specifically the LSTM architecture, as mac

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::3a5b9a013e88541ecdde1474ba1c9456
https://hal.archives-ouvertes.fr/hal-02020330/document

On Recurrent and Deep Neural Networks

Autor: Pascanu, Razvan

L'apprentissage profond est un domaine de recherche en forte croissance en apprentissage automatique qui est parvenu à des résultats impressionnants dans différentes tâches allant de la classification d'images à la parole, en passant par la mod�

Externí odkaz: http://hdl.handle.net/1866/11452