| Titre : |
Deep Learning for Physics Research |
| Type de document : |
texte imprimé |
| Auteurs : |
Martin Erdmann, Auteur ; Jonas Glombitza, Auteur ; Gregor Kasieczka, Auteur ; Uwe Klemradt, Auteur |
| Editeur : |
Singapore : World scientific publishing |
| Année de publication : |
2021 |
| Importance : |
XI-327 P. |
| Présentation : |
Relie, couv. illus. en coul : Graph. |
| Format : |
23 cm |
| ISBN/ISSN/EAN : |
978-981-1237454-- |
| Langues : |
Anglais (eng) |
| Catégories : |
(02.60) Analyse numerique et informatique
|
| Mots-clés : |
Education |
| Index. décimale : |
02.60 |
| Résumé : |
1 Scope of this textbook. - 2 Models for data analysis. - 3 Building blocks of neural networks. - 4 Optimization of network parameters. - 5 Mastering model building. - 6 Revisiting the terminology. - 7 Fully-connected networks: improving the classic all-rounder. - 8 Convolutional neural networks and analysis of image-like data. - 9 Recurrent neural networks: time series and variable input. - 10 Graph networks and convolutions beyond Euclidean domains. - 11 Multi-task learning, hybrid architectures, and operational reality. - 12. - 13Interpretability. - 13 Uncertainties and robustness. - 14 Revisiting objective functions. - 15 Beyond supervised learning. - 16 Weakly-supervised class Model independent detection of outliers and anomaliesification. - 17 Autoencoders: finding and compressing structures in data. - 18 Generative models: data from noise. - 19 Domain adaptation, refinement, unfolding. - 20 Beyond the scope of this textbook. - |
|  |
