Filters








381 Hits in 4.2 sec

Gaussian Process Classification for Variable Fidelity Data [article]

Nikita Klyuchnikov, Evgeny Burnaev
<span title="2019-10-19">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We provide an extension of Laplace inference for Gaussian process classification, that takes into account multi-fidelity data.  ...  In this paper we address a classification problem where two sources of labels with different levels of fidelity are available.  ...  In this work, we propose a co-kriging model for latent low-and high-fidelity functions and extend the Laplace inference algorithm for Gaussian process classification to handle this case.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1809.05143v3">arXiv:1809.05143v3</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/hveqcpi3evcdblfjxaumfbeeie">fatcat:hveqcpi3evcdblfjxaumfbeeie</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200826204416/https://arxiv.org/pdf/1809.05143v3.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/7b/0f/7b0f36aa0ec896e4cecfe25610fd9bafea058ab6.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1809.05143v3" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Multi-fidelity classification using Gaussian processes: accelerating the prediction of large-scale computational models [article]

Francisco Sahli Costabal, Paris Perdikaris, Ellen Kuhl, Daniel E. Hurtado
<span title="2019-05-09">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We postulate an autoregressive model between the different levels of fidelity with Gaussian process priors.  ...  We test these multi-fidelity classifiers against their single-fidelity counterpart with synthetic data, showing a median computational cost reduction of 23% for a target accuracy of 90%.  ...  This publication has received funding from Millenium Science Initiative of the Ministry of Economy, Development and Tourism of Chile, grant Nucleus for Cardiovascular Magnetic Resonance.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1905.03406v1">arXiv:1905.03406v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/z3esjyrygndbzppms5lapgjbr4">fatcat:z3esjyrygndbzppms5lapgjbr4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200825024740/https://arxiv.org/pdf/1905.03406v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/10/1d/101d3574d9c2a50fc291ef397ef3f29358aa43eb.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1905.03406v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Fast characterization of inducible regions of atrial fibrillation models with multi-fidelity Gaussian process classification [article]

Lia Gander, Simone Pezzuto, Ali Gharaviri, Rolf Krause, Paris Perdikaris, Francisco Sahli Costabal
<span title="2021-12-16">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In this work, we propose to use multi-fidelity Gaussian process classification on Riemannian manifolds to efficiently determine the regions in the atria where arrhythmias are inducible.  ...  When trained with 40 samples, our multi-fidelity classifier shows a balanced accuracy that is 10% higher than a nearest neighbor classifier used as a baseline atrial fibrillation model, and 9% higher in  ...  of this multi-fidelity Gaussian processes model.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2112.08075v2">arXiv:2112.08075v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/p5qxa2l7orarxoge5hscjkbj24">fatcat:p5qxa2l7orarxoge5hscjkbj24</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20211223161359/https://arxiv.org/pdf/2112.08075v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/73/8d/738d25a269f6d3a99bb75bc7ebec09d162f516ad.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2112.08075v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Fast Characterization of Inducible Regions of Atrial Fibrillation Models With Multi-Fidelity Gaussian Process Classification

Lia Gander, Simone Pezzuto, Ali Gharaviri, Rolf Krause, Paris Perdikaris, Francisco Sahli Costabal
<span title="2022-03-07">2022</span> <i title="Frontiers Media SA"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/evju7ot335gbrctgmomg4gg6je" style="color: black;">Frontiers in Physiology</a> </i> &nbsp;
In this work, we propose to use multi-fidelity Gaussian process classification on Riemannian manifolds to efficiently determine the regions in the atria where arrhythmias are inducible.  ...  When trained with 40 samples, our multi-fidelity classifier that combines low and high resolution models, shows a balanced accuracy that is, on average, 5.7% higher than a nearest neighbor classifier.  ...  FS and PP: formulated and implemented the classifier, whereas LG: implemented all necessary steps to perform multi-fidelity simulations on the supercomputer.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.3389/fphys.2022.757159">doi:10.3389/fphys.2022.757159</a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pubmed/35330935">pmid:35330935</a> <a target="_blank" rel="external noopener" href="https://pubmed.ncbi.nlm.nih.gov/PMC8940533/">pmcid:PMC8940533</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/h5534ukfxnddbohifmzrkmrnzy">fatcat:h5534ukfxnddbohifmzrkmrnzy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220309141812/https://fjfsdata01prod.blob.core.windows.net/articles/files/757159/pubmed-zip/.versions/2/.package-entries/fphys-13-757159-r1/fphys-13-757159.pdf?sv=2018-03-28&amp;sr=b&amp;sig=o9L1uxFRs1eSXwnxvpjz%2BV5k40lAkNb4Jq%2BGDgMoY%2B4%3D&amp;se=2022-03-09T14%3A18%3A41Z&amp;sp=r&amp;rscd=attachment%3B%20filename%2A%3DUTF-8%27%27fphys-13-757159.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/a2/f8/a2f8a95e638bfbcb49d68528d7a3d3a52fe7ca49.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.3389/fphys.2022.757159"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> frontiersin.org </button> </a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8940533" title="pubmed link"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> pubmed.gov </button> </a>

Multifidelity Bayesian Optimization for Binomial Output [article]

Leonid Matyushin, Alexey Zaytsev, Oleg Alenkin, Andrey Ustuzhanin
<span title="2019-02-19">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
The most common Gaussian process-based surrogate model assumes that the target with fixed parameters is a realization of a Gaussian process.  ...  We propose a general Gaussian process model that takes into account Bernoulli outputs.  ...  The same problem arises when you try to adapt the Gaussian processes for the task of classification [7] or robust regression with Laplace or Cauchy likelihood [8] .  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1902.06937v1">arXiv:1902.06937v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/xokhac7rtffjjhvzdnvrdy45s4">fatcat:xokhac7rtffjjhvzdnvrdy45s4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191019033513/https://arxiv.org/pdf/1902.06937v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/6a/d4/6ad4dc0ab94a52e61600d8e5425cd8a76cb4c3eb.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1902.06937v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Pillar Networks++: Distributed non-parametric deep and wide networks [article]

Biswa Sengupta, Yu Qian
<span title="2017-08-18">2017</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In the present work, we show that combining distributed Gaussian Processes with multi-stream deep convolutional neural networks (CNN) alleviate the need to augment a neural network with hand-crafted features  ...  In recent work, it was shown that combining multi-kernel based support vector machines (SVMs) can lead to near state-of-the-art performance on an action recognition dataset (HMDB-51 dataset).  ...  For inference, we have limited our experiments to the Laplace approximation inference under a distributed GP framework.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1708.06250v1">arXiv:1708.06250v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/vmucv2mdmjcezarbs63y2numye">fatcat:vmucv2mdmjcezarbs63y2numye</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191025015235/https://arxiv.org/pdf/1708.06250v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/40/09/400968c59504bb017adbc039c1738574e6e50f46.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1708.06250v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Robust Bayesian Optimization with Student-t Likelihood [article]

Ruben Martinez-Cantin, Michael McCourt, Kevin Tee
<span title="2017-07-18">2017</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
The efficiency is achieved in a similar fashion to the learning to learn methods: surrogate models (typically in the form of Gaussian processes) learn the target function and perform intelligent sampling  ...  Bayesian optimization has recently attracted the attention of the automatic machine learning community for its excellent results in hyperparameter tuning.  ...  Atkeson for releasing the code of the robot simulator and controller.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1707.05729v1">arXiv:1707.05729v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/dqnsrkr5qne2tbjfmecpe65u7u">fatcat:dqnsrkr5qne2tbjfmecpe65u7u</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200929083449/https://arxiv.org/pdf/1707.05729v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/70/0c/700c61f527458852d46ed4237fa55ec559d29256.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1707.05729v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Being Bayesian, Even Just a Bit, Fixes Overconfidence in ReLU Networks [article]

Agustinus Kristiadi, Matthias Hein, Philipp Hennig
<span title="2020-07-17">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
This indicates that a sufficient condition for a calibrated uncertainty on a ReLU network is "to be a bit Bayesian".  ...  These theoretical results validate the usage of last-layer Bayesian approximation and motivate a range of a fidelity-cost trade-off.  ...  AK is grateful to Alexander Meinke for the pre-trained models and the International Max Planck Research School for Intelligent Systems (IMPRS-IS) for support.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2002.10118v2">arXiv:2002.10118v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/6skv3cbq5bcxdelcaxgythhqqy">fatcat:6skv3cbq5bcxdelcaxgythhqqy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200722195511/https://arxiv.org/pdf/2002.10118v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/fe/21/fe21a46ac6ba41059d0ed62aaa7630540e20484a.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2002.10118v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Variational Inference in Nonconjugate Models [article]

Chong Wang, David M. Blei
<span title="2013-03-12">2013</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In this paper, we develop two generic methods for nonconjugate models, Laplace variational inference and delta method variational inference.  ...  Mean-field variational methods are widely used for approximate posterior inference in many probabilistic models.  ...  Further, we found that Laplace variational inference usually outperforms delta method variational inference, both in terms of computation time and fidelity of the approximate posterior.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1209.4360v4">arXiv:1209.4360v4</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/6lgnxdwa2nbhvghzt5agdh5phe">fatcat:6lgnxdwa2nbhvghzt5agdh5phe</a> </span>
<a target="_blank" rel="noopener" href="https://archive.org/download/arxiv-1209.4360/1209.4360.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> File Archive [PDF] </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1209.4360v4" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Omics, big data and machine learning as tools to propel understanding of biological mechanisms and to discover novel diagnostics and therapeutics

Nikolaos Perakakis, Alireza Yazdani, George E. Karniadakis, Christos Mantzoros
<span title="2018-08-08">2018</span> <i title="Elsevier BV"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/bftdqrp4bbclpnwlxeeolzbtqq" style="color: black;">Metabolism: Clinical and Experimental</a> </i> &nbsp;
modeling that can be implemented both in terms of neural networks (for classification) as well as Gaussian processes (for regression).  ...  Fig. 3 . 3 Linear (AR1) vs. non-linear (NARGP) Gaussian process regression: (a) Exact low-(red) and high-fidelity (blue) functions along with the observations used for training the multi-fidelity GP models  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1016/j.metabol.2018.08.002">doi:10.1016/j.metabol.2018.08.002</a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pubmed/30098323">pmid:30098323</a> <a target="_blank" rel="external noopener" href="https://pubmed.ncbi.nlm.nih.gov/PMC6325641/">pmcid:PMC6325641</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/m3wt2q3bc5herd27ir3eggh72y">fatcat:m3wt2q3bc5herd27ir3eggh72y</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200504134751/http://europepmc.org/backend/ptpmcrender.fcgi?accid=PMC6325641&amp;blobtype=pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/2c/b1/2cb1dd2abf6d1648660802814950055c37e5c9d7.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1016/j.metabol.2018.08.002"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> elsevier.com </button> </a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6325641" title="pubmed link"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> pubmed.gov </button> </a>

Fast Predictive Uncertainty for Classification with Bayesian Deep Networks [article]

Marius Hobbhahn, Agustinus Kristiadi, Philipp Hennig
<span title="2021-02-09">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In Bayesian Deep Learning, distributions over the output of classification neural networks are approximated by first constructing a Gaussian distribution over the weights, then sampling from it to receive  ...  We demonstrate the use of this Dirichlet approximation by using it to construct a lightweight uncertainty-aware output ranking for the ImageNet setup.  ...  Posterior inference In principle, the Gaussian over the weights required by the Laplace Bridge for BNNs (see Equation 8 ) can be constructed by any Gaussian approximate Bayesian methods such as variational  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2003.01227v2">arXiv:2003.01227v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ytailmcgrjapvanzyaxbvwc7qi">fatcat:ytailmcgrjapvanzyaxbvwc7qi</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200321100549/https://arxiv.org/pdf/2003.01227v1.pdf" title="fulltext PDF download [not primary version]" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <span style="color: #f43e3e;">&#10033;</span> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2003.01227v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Deep Multi-Fidelity Active Learning of High-dimensional Outputs [article]

Shibo Li, Robert M. Kirby, Shandian Zhe
<span title="2021-10-25">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We first develop a deep neural network-based multi-fidelity model for learning with high-dimensional outputs, which can flexibly, efficiently capture all kinds of complex relationships across the outputs  ...  To this end, we propose DMFAL, a Deep Multi-Fidelity Active Learning approach.  ...  Many multi-fidelity (MF) models have been proposed, while their active training approaches are lacking. These models are often based on Gaussian processes (GPs).  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2012.00901v2">arXiv:2012.00901v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/vdhd5xtptzewxmirxa2vxlmqri">fatcat:vdhd5xtptzewxmirxa2vxlmqri</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20211101234650/https://arxiv.org/pdf/2012.00901v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/13/ae/13aeb35db33afa117a3667349198c8aa293f3eba.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2012.00901v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Implicit Weight Uncertainty in Neural Networks [article]

Nick Pawlowski, Andrew Brock, Matthew C.H. Lee, Martin Rajchl, Ben Glocker
<span title="2018-05-25">2018</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We gratefully acknowledge the support of NVIDIA with the donation of one Titan X GPU for our research.  ...  Acknowledgements We like to thank Miguel Jacques, Pierre Richemond, Elliot Crowley, and Joseph Mellor for insightful discussions and comments on the paper.  ...  We attribute this to the multiplicative nature of MNF with the underlying Gaussian distributions. BbH fits highly complex multi-modal distributions.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1711.01297v2">arXiv:1711.01297v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/iq4bxm6osfhmzhumpb7wjatvmi">fatcat:iq4bxm6osfhmzhumpb7wjatvmi</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200911064014/https://arxiv.org/pdf/1711.01297v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/14/8e/148e2432cdf57bdac01a8592fe87a1cbb1fff101.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1711.01297v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Metric Gaussian Variational Inference [article]

Jakob Knollmüller, Torsten A. Enßlin
<span title="2020-01-30">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We propose Metric Gaussian Variational Inference (MGVI) as a method that goes beyond mean-field.  ...  The usage of natural gradient descent allows for rapid convergence.  ...  Acknowledgments We acknowledge Philipp Arras, Philipp Frank, Maksim Greiner, Sebastian Hutschenreuter, Reimar Leike, Daniel Pumpe, Martin Reinecke and Theo Steininger for fruitful discussions.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1901.11033v3">arXiv:1901.11033v3</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/4xth43f4mzaanir4rwr5hufq2i">fatcat:4xth43f4mzaanir4rwr5hufq2i</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200321040503/https://arxiv.org/pdf/1901.11033v3.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1901.11033v3" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Multi-task Learning for Aggregated Data using Gaussian Processes [article]

Fariba Yousefi, Michael Thomas Smith, Mauricio A. Álvarez
<span title="2020-02-19">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In this paper, we present a novel multi-task learning model based on Gaussian processes for joint learning of variables that have been aggregated at different input scales.  ...  For example, census data for a population is usually given as averages defined over time periods or spatial resolutions (cities, regions or countries).  ...  We appeal to the flexibility of Gaussian processes (GPs) for developing a prior over such type of datasets and we also provide a scalable approach for variational Bayesian inference.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1906.09412v4">arXiv:1906.09412v4</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/3oek4omdvnhd7g3eqb7ghxoxsi">fatcat:3oek4omdvnhd7g3eqb7ghxoxsi</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200321094039/https://arxiv.org/pdf/1906.09412v4.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1906.09412v4" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>
&laquo; Previous Showing results 1 &mdash; 15 out of 381 results