LBBNN

LBBNN-package

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) tools:::Rd_expr_doi("10.3390/math12060788"), using the local reparametrization trick as in Skaaret-Lund et al. (2024) <a href="https://openreview.net/pdf?id=d6kqUKzG3V">https://openreview.net/pdf?id=d6kqUKzG3V</a>. Input-skip connections are also supported, as described in Høyheim et al. (2025) tools:::Rd_expr_doi("10.48550/arXiv.2503.10496").

internal

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>.

Lars Skaaret-Lund

Latent Binary Bayesian Neural Networks Using 'torch'

Aliaksandr Hubin

Eirik Høyheim

LBBNN-package function

Maintainer: Lars Skaaret-Lund <a href="/link/lars.skaaret-lund%40nmbu.no?package=LBBNN&version=0.1.2" data-mini-rdoc="LBBNN::lars.skaaret-lund@nmbu.no">lars.skaaret-lund@nmbu.no</a>
Authors:<ul>
<li>Aliaksandr Hubin <a href="/link/aliaksandr.hubin%40nmbu.no?package=LBBNN&version=0.1.2" data-mini-rdoc="LBBNN::aliaksandr.hubin@nmbu.no">aliaksandr.hubin@nmbu.no</a></li>
<li>Eirik Høyheim <a href="/link/eirik.hoyheim%40ffi.no?package=LBBNN&version=0.1.2" data-mini-rdoc="LBBNN::eirik.hoyheim@ffi.no">eirik.hoyheim@ffi.no</a></li>
</ul>

Author

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) tools:::Rd_expr_doi("10.3390/math12060788"), using the local reparametrization trick as in Skaaret-Lund et al. (2024) <a href='https://openreview.net/pdf?id=d6kqUKzG3V'>https://openreview.net/pdf?id=d6kqUKzG3V</a>. Input-skip connections are also supported, as described in Høyheim et al. (2025) tools:::Rd_expr_doi("10.48550/arXiv.2503.10496").

LBBNN: Latent Binary Bayesian Neural Networks Using 'torch' — LBBNN-package

Maintainer: Lars Skaaret-Lund <a href='mailto:lars.skaaret-lund@nmbu.no'>lars.skaaret-lund@nmbu.no</a>
Authors:<ul>
<li>Aliaksandr Hubin <a href='mailto:aliaksandr.hubin@nmbu.no'>aliaksandr.hubin@nmbu.no</a></li>
<li>Eirik Høyheim <a href='mailto:eirik.hoyheim@ffi.no'>eirik.hoyheim@ffi.no</a></li>
</ul>

LBBNN-package: LBBNN: Latent Binary Bayesian Neural Networks Using 'torch'

Description

Arguments

Author