Mean Field Bayes Backpropagation: scalable training of multilayer neural networks with binary weights

Soudry, Daniel; Meir, Ron

doi:10.48550/arxiv.1310.1867

Cited by 2 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this framework, the optimal weight is equivalent to minimizing the objective function. For details of Bayesian regularization, interested readers can see different studies (Burden & Winkler, 2009;Duhoux et al, 2001;Foresee & Hagan, 1997;Ghosh & Doshi-Velez, 2017;Hernández-Lobato & Adams, 2015;Khan & Coulibaly, 2006;MacKay, 1992;Soudry & Meir, 2013;Tebabal et al, 2019). 5for each hour of measurements, is taken to be the Id or Az of the hour of the day.…”

Section: Neural Network (Nn)mentioning

confidence: 99%

Comparative Assessment of the Performances of GPS‐TEC Assisted NTCM, NeQuick2 and Neural Network Models to Describe the East‐African Equatorial Ionosphere

2023

View full text Add to dashboard Cite

The ionosphere is part of the upper atmosphere of Earth where EUV, X-ray radiations, and energetic particles from the sun ionize the atoms and molecules. It is a region where free electrons occur in a high enough density to influence the propagation of radio waves with measurable effects. It is one of the highly variable regions of the Earth-upper atmosphere. Nowadays, society has become reliant on systems that work with communication of the satellites whose performance is dependent on the state of the ionosphere. The ionosphere causes a significant error in single-frequency GPS navigation (Davies, 1990;Tiwari et al., 2013). The temporal and spatial variation of electron density (total electron content, TEC) of the ionosphere affects the GPS signal propagation by introducing delay (slowing down and bending propagation) and scintillation (Jakowski et al., 2012;Opperman et al., 2007;Tulunay et al., 2004Tulunay et al., , 2006. Understanding the ionosphere dynamics and developing a model for nowcasting and forecasting its TEC might mitigate the ionosphere effect. In this regard, the interest in specifying the ionosphere TEC is not only for climatic-like situations but also for weather-like behavior.Global empirical models such as IRI (Bilitza, 2018), NeQuick2 (Nava et al., 2008), and NTCM (Jakowski et al., 2011) are good in capturing the median behavior of the ionosphere. These models have been designed for scientific study and radio communication applications. For example, NeQUick2 is adopted for TEC prediction to improve communication quality for International Telecommunication Union Radiocommunication Sector (ITU-R). It is also approved to be used for single frequency ionospheric error correction in the framework of GALILEO (Nigussie et al., 2012 and references therein). These models have shown different performance in their standard application in describing the observed TEC; for example, Nigussie et al. ( 2013) have compared the performance of NeQuick2 and IRI-2007 models in describing the East-African ionospheric TEC and found, for both models, the modeled and experimental VTEC has shown significant discrepancy. The performance of NTCM in its standard applications has shown also a significant discrepancy to the experimental VTEC (Getahun & Nigussie, 2021). The weak performances of these models in their standard applications indicate limitations in reproducing and forecasting the local and instantaneous conditions, especially in regions of unique irregularities are common (Scherliess et al., 2004;Sojka et al., 2007).

show abstract

Section: Neural Network (Nn)mentioning

confidence: 99%

Comparative Assessment of the Performances of GPS‐TEC Assisted NTCM, NeQuick2 and Neural Network Models to Describe the East‐African Equatorial Ionosphere

2023

View full text Add to dashboard Cite

show abstract

“…For details of Bayesian regularization, interested readers can see different studies (Foresee and Hagan, 1997;MacKay, 1992;Duhoux et al, 2001;Khan and Coulibaly, 2006;Burden and Winkler, 2009;Soudry and Meir, 2013;Hernández-Lobato and Adams, 2015;Ghosh and Doshi-Velez,2017;Tebabal et al2019 ).…”

Section: Neural Network (Nn)mentioning

confidence: 99%

Comparative assessment of the performances of GPS-TEC assisted NTCM, NeQuick2 and Neural Network models to describe the East-African equatorial Ionosphere

Gelaw

Nigussie

Tebabal³

2022

Preprint

View full text Add to dashboard Cite

Different ionospheric climatological models such as NeQuick2 and NTCM have been developed to mitigate the ionosphere impact on the trans-ionosphere propagating radio wave. Moreover, the Neural Network (NN) is used to model and characterize the ionosphere. However, no one has compared the performances of NeQuick2, NTCM, and NN after adapting to GPS TEC.This study evaluates their performances in the East-African region in 2013 and 2015. It has been done by computing their drivers (effective ionization level, Az for NeQuick2 and ionization driving index, Id for NTCM) through least-square fitting to TEC observation sense. NN-algorithm has also been trained and tested with observed TEC used for NeQuick2 and NTCM adaptation. The annual performance test has shown that the correlation coefficient (R) values between observed and NTCM modeled TEC, after an adaption, are better than the corresponding values obtained from NeQuick2 and NN. It also shows that the standard deviations (STD) and root-mean-square errors (RMSE) obtained for NTCM are smaller than the STD and RMSE computed for NeQuick2 and NN. On the other hand, the daily performance test of now-casting and predicting showed that the NN performs the best, followed by NTCM. However, the 1-hour prediction test showed that NTCM performs the best among the models considered in this study. Hosted fileessoar.10512098.1.docx available at https://authorea.com/users/524617/articles/595516comparative-assessment-of-the-performances-of-gps-tec-assisted-ntcm-nequick2-and-neuralnetwork-models-to-describe-the-east-african-equatorial-ionosphere Comparative assessment of the performances of GPS-TEC assisted NTCM, NeQuick2 and Neural Network models to describe the East-African equatorial Ionosphere

show abstract

Mean Field Bayes Backpropagation: scalable training of multilayer neural networks with binary weights

Cited by 2 publications

References 17 publications

Comparative Assessment of the Performances of GPS‐TEC Assisted NTCM, NeQuick2 and Neural Network Models to Describe the East‐African Equatorial Ionosphere

Comparative Assessment of the Performances of GPS‐TEC Assisted NTCM, NeQuick2 and Neural Network Models to Describe the East‐African Equatorial Ionosphere

Comparative assessment of the performances of GPS-TEC assisted NTCM, NeQuick2 and Neural Network models to describe the East-African equatorial Ionosphere

Contact Info

Product

Resources

About