https://larselund.r-universe.devRecent package updates in larselundR-universehttps://github.com/larselund.pnghttps://larselund.r-universe.devWed, 01 Jul 2026 09:39:48 GMTlars.skaaret-lund@nmbu.no (Lars Skaaret-Lund)Latent binary Bayesian neural networks (LBBNNs) are implemented using 'torch', an R interface to the LibTorch backend. Supports mean-field variational inference as well as flexible variational posteriors using normalizing flows. The standard LBBNN implementation follows Hubin and Storvik (2024) <doi:10.3390/math12060788>, using the local reparametrization trick as in Skaaret-Lund et al. (2024) <https://openreview.net/pdf?id=d6kqUKzG3V>. Input-skip connections are also supported, as described in Høyheim et al. (2025) <doi:10.48550/arXiv.2503.10496>.https://github.com/r-universe/larselund/actions/runs/28513584300Wed, 01 Jul 2026 09:39:48 GMTLBBNN0.1.6successhttps://larselund.r-universe.devhttps://github.com/larselund/lbbnnconvolutional_architecture.Rmdconvolutional_architecture.htmlclassification: convolutional architecture2026-06-20 11:41:132026-06-30 12:19:30small_dataset_classification.Rmdsmall_dataset_classification.htmlclassification: gallstone dataset2026-06-20 11:41:132026-06-30 12:19:30getting_started.Rmdgetting_started.htmlGetting started with LBBNN2026-06-20 11:41:132026-06-30 12:19:30variable_selection_linear_data.Rmdvariable_selection_linear_data.htmlvariable selection: linear-data2026-06-20 11:41:132026-06-30 12:19:30variable_selection_non_linear_data.Rmdvariable_selection_non_linear_data.htmlvariable selection: non-linear data2026-06-20 11:41:132026-06-30 12:19:30