Centre dedocumentation
scientifique

Volcano-seismic transfer learning and uncertainty quantification with bayesian neural networks / Angel Bueno in IEEE Transactions on geoscience and remote sensing, vol 58 n° 2 (February 2020)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 58 n° 2 (February 2020) . - pp Titre : Volcano-seismic transfer learning and uncertainty quantification with bayesian neural networks Type de document : Article/Communication Auteurs : Angel Bueno, Auteur ; Carmen Benitez, Auteur ; Silvio De Angelis, Auteur ; et al., Auteur Année de publication : 2020 Article en page(s) : pp Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Statistiques [Termes descripteurs IGN] apprentissage profond [Termes descripteurs IGN] classification bayesienne [Termes descripteurs IGN] classification par réseau neuronal [Termes descripteurs IGN] forme d'onde [Termes descripteurs IGN] incertitude des données [Termes descripteurs IGN] réseau bayesien [Termes descripteurs IGN] réseau de neurones profond [Termes descripteurs IGN] Russie [Termes descripteurs IGN] séisme [Termes descripteurs IGN] sismologie [Termes descripteurs IGN] surveillance géologique [Termes descripteurs IGN] volcanologie [Termes descripteurs IGN] Washington (Etats-Unis ; état) Résumé : (auteur) Over the past few years, deep learning (DL) has emerged as an important tool in the fields of volcano and earthquake seismology. However, these methods have been applied without performing thorough analyses of the associated uncertainties. Here, we propose a solution to enhance volcano-seismic monitoring systems, through probabilistic Bayesian DL; we implement and demonstrate a workflow for waveform classification, rapid quantification of the associated uncertainty, and link these uncertainties to changes in volcanic unrest. Specifically, we introduce Bayesian neural networks (BNNs) to perform event identification, classification, and their estimated uncertainty on data gathered at two active volcanoes, Mount St. Helens, Washington, USA, and Bezymianny, Kamchatka, Russia. We demonstrate how BNNs achieve excellent performance (92.08%) in discriminating both the type of event and its origin when the two data sets are merged together, and no additional training information is provided. Finally, we demonstrate that the data representations learned by the BNNs are transferable across different eruptive periods. We also find that the estimated uncertainty is related to changes in the state of unrest at the volcanoes and propose that it could be used to gauge whether the learned models may be exported to other eruptive scenarios. Numéro de notice : A2020-094 Thématique : MATHEMATIQUE/POSITIONNEMENT Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2019.2941494 date de publication en ligne : 07/10/2019 En ligne : https://doi.org/10.1109/TGRS.2019.2941494 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=94657 [article]

A Bayesian network-based method to alleviate the ill-posed inverse problem: A case study on leaf area index and canopy water content retrieval / Xingwen Quan in IEEE Transactions on geoscience and remote sensing, vol 53 n° 12 (December 2015)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 53 n° 12 (December 2015) . - pp 6507 - 6517 Titre : A Bayesian network-based method to alleviate the ill-posed inverse problem: A case study on leaf area index and canopy water content retrieval Type de document : Article/Communication Auteurs : Xingwen Quan, Auteur ; Binbin He, Auteur ; Xing Li, Auteur Année de publication : 2015 Article en page(s) : pp 6507 - 6517 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications de télédétection [Termes descripteurs IGN] corrélation d'images [Termes descripteurs IGN] image Landsat-8 [Termes descripteurs IGN] Leaf Area Index [Termes descripteurs IGN] probabilité [Termes descripteurs IGN] problème inverse [Termes descripteurs IGN] réseau bayesien [Termes descripteurs IGN] teneur en eau de la végétation Résumé : (auteur) Retrieval of vegetation parameters from remotely sensed data using a radiative transfer model is generally hampered by the ill-posed inverse problem, which dramatically decreases the precision level of retrieved parameters. The purpose of this study was to use a Bayesian network-based method to allow the alleviation of the ill-posed inverse problem. This was achieved by introducing the correlations between the model free parameters into their prior joint probability distribution (PJPD), allowing the reduction of the probabilities of unrealistic combinations. Three sampling strategies intended to design three types of PJPDs that considered different correlations (represented by a correlation matrix) were presented. They were multivariate uniform distribution composed by independent free parameters, multivariate uniform distribution based on a simple correlation matrix, and multivariate Gaussian distribution based on a complicated correlation matrix, respectively. A case study of the presented method to retrieve leaf area index (LAI) and canopy water content (CWC) using the PROSAIL_5B (PROSPECT-5 + 4SAIL) model from Landsat 8 products was implemented. Results indicate that the presented method greatly improves the precision level of target parameters, with the coefficient of determination R2 of 0.69, 0.77, and 0.82 and root-mean-square error (RMSE) of 0.55, 0.51, and 0.44 m2 · m-2 for LAI and R2 = 0.68, 0.78, and 0.84 and RMSE = 230, 198, and 166 g · m-2 for CWC, respectively. Hence, the ill-posed inverse problem can be alleviated by the presented method, which can be widely applied for vegetation parameters retrieval. Numéro de notice : A2015-838 Thématique : FORET/IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2015.2442999 date de publication en ligne : 30/06/2015 En ligne : https://doi.org/10.1109/TGRS.2015.2442999 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=79172 [article]

Réservation

Réserver ce document

Exemplaires (1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
065-2015121	SL	Revue	Centre de documentation	Revues en salle	Disponible

Bayesian belief networks as a versatile method for assessing uncertainty in land-change modeling / Carsten Krüger in International journal of geographical information science IJGIS, vol 29 n° 1 (January 2015)  

Public [article]  inInternational journal of geographical information science IJGIS > vol 29 n° 1 (January 2015) . - pp 111 - 131 Titre : Bayesian belief networks as a versatile method for assessing uncertainty in land-change modeling Type de document : Article/Communication Auteurs : Carsten Krüger, Auteur ; Tobia Lakes, Auteur Année de publication : 2015 Article en page(s) : pp 111 - 131 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Analyse spatiale [Termes descripteurs IGN] aide à la décision [Termes descripteurs IGN] Amazonie [Termes descripteurs IGN] Brésil [Termes descripteurs IGN] changement d'occupation du sol [Termes descripteurs IGN] changement d'utilisation du sol [Termes descripteurs IGN] déboisement [Termes descripteurs IGN] incertitude des données [Termes descripteurs IGN] modèle conceptuel de données spatio-temporelles [Termes descripteurs IGN] réseau bayesien [Termes descripteurs IGN] théorie de Dempster-Shafer Résumé : (auteur) Land-use and land-cover change modeling helps us to understand the driving factors and impacts of human-induced land changes better, and depict likely future development paths. Uncertainty associated with various steps in the modeling process substantially influences the reliability of the results, but until now it has only rarely been addressed. In this study, we explore uncertainty in land-change modeling using a probabilistic approach based on Bayesian belief networks. We apply this approach to a case study of deforestation in the Brazilian Amazon and identify three modeling steps as sources of uncertainty: model structure, variable selection, and data preprocessing. For these three steps, we quantify the uncertainty and the respective impact on the outcome accuracy. The results indicate remarkable uncertainties in each of the steps. We demonstrate that a higher uncertainty in the land-change modeling process does not necessarily lead to a lower accuracy of the modeling outcome. Moreover, we show that the different uncertainty sources only slightly influence the ratio between quantity disagreement and allocation disagreement for the modeling outcome. We conclude that uncertainty is inherent in land-change modeling, and that future studies should address this uncertainty more explicitly to improve the robustness of modeling outcomes for science and decision-making. Numéro de notice : A2015-532 Thématique : GEOMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1080/13658816.2014.949265 En ligne : http://www.tandfonline.com/doi/full/10.1080/13658816.2014.949265 Format de la ressource électronique : URL bulletin Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=77763 [article]

The Bayesian detection of discontinuities in a polynomial regression and its application to the cycle-slip problem / M.C. DE Lacy in Journal of geodesy, vol 82 n° 9 (September 2008)

Public [article]  inJournal of geodesy > vol 82 n° 9 (September 2008) . - pp 527 - 542 Titre : The Bayesian detection of discontinuities in a polynomial regression and its application to the cycle-slip problem Type de document : Article/Communication Auteurs : M.C. DE Lacy, Auteur ; M. Reguzzoni, Auteur ; Fernando Sanso, Auteur ; Giovanna Venuti, Auteur Année de publication : 2008 Article en page(s) : pp 527 - 542 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Géodésie spatiale [Termes descripteurs IGN] Bernese [Termes descripteurs IGN] correction du signal [Termes descripteurs IGN] phase GNSS [Termes descripteurs IGN] Ploynôme [Termes descripteurs IGN] régression [Termes descripteurs IGN] réseau bayesien [Termes descripteurs IGN] traitement de données GNSS [Termes descripteurs IGN] valeur discrète Résumé : (Auteur) This paper deals with the problem of detecting and correcting cycle-slips in Global Navigation Satellite System (GNSS) phase data by exploiting the Bayesian theory. The method is here applied to undifferenced observations, because repairing cycle-slips already at this stage could be a useful pre-processing tool, especially for a network of permanent GNSS stations. If a dual frequency receiver is available, the cycle-slips can be easily detected by combining two phase observations or phase and range observations from a single satellite to a single receiver. These combinations, expressed in a distance unit form, are completely free from the geometry and depend only on the ionospheric effect, on the electronic biases and on the initial integer ambiguities; since these terms are expected to be smooth in time, at least in a short period, a cycle-slip in one or both the two carriers can be modelled as a discontinuity in a polynomial regression. The proposed method consists in applying the Bayesian theory to compute the marginal posterior distribution of the discontinuity epoch and to detect it as a maximum a posteriori (MAP) in a very accurate way. Concerning the cycle-slip correction, a couple of simultaneous integer slips in the two carriers is chosen by maximazing the conditional posterior distribution of the discontinuity amplitude given the detected epoch. Numerical experiments on simulated and real data show that the discontinuities with an amplitude 2 or 3 times larger than the noise standard deviation are successfully identified. This means that the Bayesian approach is able to detect and correct cycle-slips using undifferenced GNSS observations even if the slip occurs by one cycle. A comparison with the scientific software BERNESE 5.0 confirms the good performance of the proposed method, especially when data sampled at high frequency (e.g. every 1 s or every 5 s) are available. Copyright Springer Numéro de notice : A2008-349 Thématique : POSITIONNEMENT Nature : Article nature-HAL : ArtAvecCL-RevueIntern Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=29342 [article]

Réservation

Réserver ce document

Exemplaires (2)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
266-08081	RAB	Revue	Centre de documentation	En réserve 3L	Disponible
266-08082	RAB	Revue	Centre de documentation	En réserve 3L	Disponible

Semantic-sensitive satellite image retrieval / Y. Li in IEEE Transactions on geoscience and remote sensing, vol 45 n° 4 (April 2007)

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 45 n° 4 (April 2007) . - pp 853 - 860 Titre : Semantic-sensitive satellite image retrieval Type de document : Article/Communication Auteurs : Y. Li, Auteur ; T.R. Bretschneider, Auteur Année de publication : 2007 Article en page(s) : pp 853 - 860 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Bases de données localisées [Termes descripteurs IGN] base de données d'images [Termes descripteurs IGN] image satellite [Termes descripteurs IGN] inférence sémantique [Termes descripteurs IGN] recherche d'information [Termes descripteurs IGN] réseau bayesien Résumé : (Auteur) Content-based image-retrieval techniques based on query scenes are a powerful means for exploration and mining of large remote sensing image databases. However, the gap between low-level unsupervised extracted features in content-based retrieval and the high-level semantic concepts of user queries limits the performance. Therefore, this paper proposes a specialized approach using a context-sensitive Bayesian network for semantic inference of segmented scenes. The regions' remote sensing related semantic concepts are inferred in a multistage process based on their spectral and textural characteristics as well as the semantics of adjacent regions. During the actual retrieval, the semantics are employed for the extraction of candidate scenes which are evaluated and ranked in a consecutive step. The approach was implemented and compared with a different strategy that utilizes the extracted features from the imagery directly to infer the semantics. In summary, the developed system achieved higher precision and recall rates using the same training data Copyright IEEE Numéro de notice : A2007-219 Thématique : GEOMATIQUE/INFORMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=28582 [article]

Réservation

Réserver ce document

Exemplaires (1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
065-07041	RAB	Revue	Centre de documentation	En réserve 3L	Disponible

Optimisation en traitement du signal et de l'image / P. Siarry (2007)

Permalink

Application des réseaux bayésiens de classification dans les systèmes d'informatin géographique / Marie-Aline Cavarroc (2005)

Permalink

Hierarchical Bayesian nets for building extraction using dense digital surface models / A. Brunn in ISPRS Journal of photogrammetry and remote sensing, vol 53 n° 5 (September - October 1998)

Permalink

Révision d'information dans un SIG / Marie-Aline Cavarroc (1998)

Permalink

Special interim technical report / Robert M. Haralick (1969)

Permalink