Filters








311 Hits in 2.3 sec

Bayesian Variable Selection for Globally Sparse Probabilistic PCA [article]

Charles Bouveyron, Pierre Latouche, Pierre-Alexandre Mattei
<span title="2016-09-20">2016</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Sparse versions of principal component analysis (PCA) have imposed themselves as simple, yet powerful ways of selecting relevant features of high-dimensional data in an unsupervised manner. However, when several sparse principal components are computed, the interpretation of the selected variables is difficult since each axis has its own sparsity pattern and has to be interpreted separately. To overcome this drawback, we propose a Bayesian procedure called globally sparse probabilistic PCA
more &raquo; ... CA) that allows to obtain several sparse components with the same sparsity pattern. This allows the practitioner to identify the original variables which are relevant to describe the data. To this end, using Roweis' probabilistic interpretation of PCA and a Gaussian prior on the loading matrix, we provide the first exact computation of the marginal likelihood of a Bayesian PCA model. To avoid the drawbacks of discrete model selection, a simple relaxation of this framework is presented. It allows to find a path of models using a variational expectation-maximization algorithm. The exact marginal likelihood is then maximized over this path. This approach is illustrated on real and synthetic data sets. In particular, using unlabeled microarray data, GSPPCA infers much more relevant gene subsets than traditional sparse PCA algorithms.
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1605.05918v2">arXiv:1605.05918v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/sckpgvdblnbpviisplroaxgmma">fatcat:sckpgvdblnbpviisplroaxgmma</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200823011543/https://arxiv.org/pdf/1605.05918v2.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/e4/09/e4090838bc8065a5a496298a437d2db64d23dfc8.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1605.05918v2" 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>

Exact Dimensionality Selection for Bayesian PCA [article]

Charles Bouveyron , Pierre-Alexandre Mattei
<span title="2019-05-21">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Following Mattei (2017) , we can state the following lemma.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1703.02834v2">arXiv:1703.02834v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/yyjwsjovrfawhozgbtzd4urlre">fatcat:yyjwsjovrfawhozgbtzd4urlre</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200908101044/https://arxiv.org/pdf/1703.02834v1.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> <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/02/93/0293042e60e38ac7beece97d7ef4260d7307a8ee.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1703.02834v2" 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>

A Parsimonious Tour of Bayesian Model Uncertainty [article]

Pierre-Alexandre Mattei
<span title="2020-09-25">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Acknowledgement This review paper is an extended version of the first chapter of my PhD thesis, which was fuelled by the wonderful energy and guidance of my advisors, Charles Bouveyron and Pierre Latouche  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1902.05539v2">arXiv:1902.05539v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/fz246wnxure6ripw6zy2djmz6e">fatcat:fz246wnxure6ripw6zy2djmz6e</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20201001200032/https://arxiv.org/pdf/1902.05539v2.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/1902.05539v2" 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>

Bayesian variable selection for globally sparse probabilistic PCA

Charles Bouveyron, Pierre Latouche, Pierre-Alexandre Mattei
<span title="">2018</span> <i title="Institute of Mathematical Statistics"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/t6cjwmijl5fi5ksnplmoz7xzji" style="color: black;">Electronic Journal of Statistics</a> </i> &nbsp;
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1214/18-ejs1450">doi:10.1214/18-ejs1450</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/imv2bdsolbfpba3vh2h3yj4oty">fatcat:imv2bdsolbfpba3vh2h3yj4oty</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20190505114758/https://hal.archives-ouvertes.fr/hal-01310409/file/GSPPCAv2.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/a0/78/a078b77a4ebbf39d7e160f7291df3327c7967da2.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1214/18-ejs1450"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> Publisher / doi.org </button> </a>

Multiplying a Gaussian Matrix by a Gaussian Vector [article]

Pierre-Alexandre Mattei
<span title="2017-04-05">2017</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Acknowledgements I thank Charles Bouveyron, Pierre Latouche, Brendan Murphy and Christian Robert for fruitful advices and discussions.  ...  ) columns and a zero-mean isotropic Gaussian vector follows a symmetric multivariate generalized Laplace distribution, a result that has useful applications for Bayesian principal component analysis Mattei  ...  While the proof of Bouveyron, Latouche, and Mattei (2016) relied on characteristic functions and the properties of Bessel functions, the proof that we presented here is closer in spirit to the one of  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1702.02815v2">arXiv:1702.02815v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/omydw5lyqrfbnbwdyf5wnnlj7q">fatcat:omydw5lyqrfbnbwdyf5wnnlj7q</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200914045513/https://arxiv.org/pdf/1702.02815v2.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/49/97/4997dcf47aac428a9dd82d347760df9b40c445fa.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1702.02815v2" 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>

Unobserved classes and extra variables in high-dimensional discriminant analysis [article]

Michael Fop, Pierre-Alexandre Mattei, Charles Bouveyron, Thomas Brendan Murphy
<span title="2021-02-03">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In supervised classification problems, the test set may contain data points belonging to classes not observed in the learning phase. Moreover, the same units in the test data may be measured on a set of additional variables recorded at a subsequent stage with respect to when the learning sample was collected. In this situation, the classifier built in the learning phase needs to adapt to handle potential unknown classes and the extra dimensions. We introduce a model-based discriminant approach,
more &raquo; ... Dimension-Adaptive Mixture Discriminant Analysis (D-AMDA), which can detect unobserved classes and adapt to the increasing dimensionality. Model estimation is carried out via a full inductive approach based on an EM algorithm. The method is then embedded in a more general framework for adaptive variable selection and classification suitable for data of large dimensions. A simulation study and an artificial experiment related to classification of adulterated honey samples are used to validate the ability of the proposed framework to deal with complex situations.
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2102.01982v1">arXiv:2102.01982v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/fllesuodyjgaxd7woqgip2fd44">fatcat:fllesuodyjgaxd7woqgip2fd44</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210205092636/https://arxiv.org/pdf/2102.01982v1.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/4f/00/4f0001200e6621f6aac942b00c5395ef81831b4e.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2102.01982v1" 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>

Multiplying a Gaussian matrix by a Gaussian vector

Pierre-Alexandre Mattei
<span title="">2017</span> <i title="Elsevier BV"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/rwdjnin7wvc3xmoc6by27vqkgq" style="color: black;">Statistics and Probability Letters</a> </i> &nbsp;
Acknowledgements I thank Charles Bouveyron, Pierre Latouche, Brendan Murphy and Christian Robert for fruitful advices and discussions.  ...  ) columns and a zero-mean isotropic Gaussian vector follows a symmetric multivariate generalized Laplace distribution, a result that has useful applications for Bayesian principal component analysis Mattei  ...  While the proof of Bouveyron, Latouche, and Mattei (2016) relied on characteristic functions and the properties of Bessel functions, the proof that we presented here is closer in spirit to the one of  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1016/j.spl.2017.04.004">doi:10.1016/j.spl.2017.04.004</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/zhvb2xt2fbbphjbedf5mrk6zqu">fatcat:zhvb2xt2fbbphjbedf5mrk6zqu</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20180725010439/https://hal.archives-ouvertes.fr/hal-01462941/file/draftGL.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/dc/58/dc58ab65e4b8ffeb529df38993e9fd3a62a85784.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1016/j.spl.2017.04.004"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> elsevier.com </button> </a>

A Multi-stage deep architecture for summary generation of soccer videos [article]

Melissa Sanabria, Frédéric Precioso, Pierre-Alexandre Mattei, Thomas Menguy
<span title="2022-05-02">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Supplementary Material: A Multi-stage deep architecture for summary generation of soccer videos Melissa Sanabria, Frédéric Precioso, Pierre-Alexandre Mattei, and Thomas Menguy In this Supplementary Material  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2205.00694v1">arXiv:2205.00694v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/tjarfapwvja4vdkaba2mhwgaui">fatcat:tjarfapwvja4vdkaba2mhwgaui</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220506201859/https://arxiv.org/pdf/2205.00694v1.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/88/08/880811fc3bc386f652b7921707284666b7272808.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2205.00694v1" 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>

MIWAE: Deep Generative Modelling and Imputation of Incomplete Data [article]

Pierre-Alexandre Mattei, Jes Frellsen
<span title="2019-02-04">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Correspondence to: Pierre-Alexandre Mattei <pima@itu.dk>, Jes Frellsen <jefr@itu.dk>.  ...  ., 2018) and nonparametric mixture models (Mattei & Frellsen, 2018a) .  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1812.02633v2">arXiv:1812.02633v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/rouuxoskcndetectdxpqjyds3u">fatcat:rouuxoskcndetectdxpqjyds3u</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191021000836/https://arxiv.org/pdf/1812.02633v2.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/6f/f0/6ff092446041cb11b110fd9141a7fe847523e908.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1812.02633v2" 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>

Combining a relaxed EM algorithm with Occam's razor for Bayesian variable selection in high-dimensional regression

Pierre Latouche, Pierre-Alexandre Mattei, Charles Bouveyron, Julien Chiquet
<span title="">2016</span> <i title="Elsevier BV"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/dzig5epjvbgqpkmvs5urnktqi4" style="color: black;">Journal of Multivariate Analysis</a> </i> &nbsp;
We address the problem of Bayesian variable selection for high-dimensional linear regression. We consider a generative model that uses a spike-and-slab-like prior distribution obtained by multiplying a deterministic binary vector, which traduces the sparsity of the problem, with a random Gaussian parameter vector. The originality of the work is to consider inference through relaxing the model and using a type-II log-likelihood maximization based on an EM algorithm. Model selection is performed
more &raquo; ... fterwards relying on Occam's razor and on a path of models found by the EM algorithm. Numerical comparisons between our method, called spinyReg, and state-of-the-art high-dimensional variable selection algorithms (such as lasso, adaptive lasso, stability selection or spike-and-slab procedures) are reported. Competitive variable selection results and predictive performances are achieved on both simulated and real benchmark data sets. An original regression data set involving the prediction of the number of visitors of the Orsay museum in Paris using bike-sharing system data is also introduced, illustrating the efficiency of the proposed approach. An R package implementing the spinyReg method is currently under development and is available at https://r-forge.r-project.org/projects/spinyreg.
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1016/j.jmva.2015.09.004">doi:10.1016/j.jmva.2015.09.004</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/pg2gwwcwhna5pfmu5wnzyvg4s4">fatcat:pg2gwwcwhna5pfmu5wnzyvg4s4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20190217011014/https://core.ac.uk/download/pdf/52813630.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/02/14/0214ef736bb8bc4e8b54d6ecfd21363eee2fa21a.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1016/j.jmva.2015.09.004"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> elsevier.com </button> </a>

Molecular architecture of the PBP2–MreC core bacterial cell wall synthesis complex

Carlos Contreras-Martel, Alexandre Martins, Chantal Ecobichon, Daniel Maragno Trindade, Pierre-Jean Matteï, Samia Hicham, Pierre Hardouin, Meriem El Ghachi, Ivo G. Boneca, Andréa Dessen
<span title="2017-10-03">2017</span> <i title="Springer Nature"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/a4wan6l5o5dfzn767kyz7jqevi" style="color: black;">Nature Communications</a> </i> &nbsp;
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1038/s41467-017-00783-2">doi:10.1038/s41467-017-00783-2</a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pubmed/28974686">pmid:28974686</a> <a target="_blank" rel="external noopener" href="https://pubmed.ncbi.nlm.nih.gov/PMC5626683/">pmcid:PMC5626683</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/roomdqnkgrfo3ftlqxfjawpdde">fatcat:roomdqnkgrfo3ftlqxfjawpdde</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20180720224342/http://hal.univ-grenoble-alpes.fr/hal-01618613/document" 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/45/86/4586d82dd2677dc60d3e6e073e21d7754cb97f0d.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1038/s41467-017-00783-2"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> nature.com </button> </a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5626683" title="pubmed link"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> pubmed.gov </button> </a>

Class-specific variable selection in high-dimensional discriminant analysis through Bayesian Sparsity

Fanny Orlhac, Pierre-Alexandre Mattei, Charles Bouveyron, Nicholas Ayache
<span title="2018-11-18">2018</span> <i title="Wiley"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/mgvnagfuwjddxbu2ryhv4mxcii" style="color: black;">Journal of Chemometrics</a> </i> &nbsp;
Although the ongoing digital revolution in fields such as chemometrics, genomics or personalized medicine gives hope for considerable progress in these areas, it also provides more and more high-dimensional data to analyze and interpret. A common usual task in those fields is discriminant analysis, which however may suffer from the high dimensionality of the data. The recent advances, through subspace classification or variable selection methods, allowed to reach either excellent classification
more &raquo; ... performances or useful visualizations and interpretations. Obviously, it is of great interest to have both excellent classification accuracies and a meaningful variable selection for interpretation. This work addresses this issue by introducing a subspace discriminant analysis method which performs a class-specific variable selection through Bayesian sparsity. The resulting classification methodology is called sparse high-dimensional discriminant analysis (sHDDA). Contrary to most sparse methods which are based on the Lasso, sHDDA relies on a Bayesian modeling of the sparsity pattern and avoids the painstaking and sensitive cross-validation of the sparsity level. The main features of sHDDA are illustrated on simulated and real-world data. In particular, we propose an exemplar application to cancer characterization based on medical imaging using radiomic feature extraction is in particular proposed. Sparse HDDA This section introduces the model for sparse discriminant analysis and discusses model inference. This model extends the HDDA model of Bouveyron et al. (2007) by incorporating a sparsity pattern for each class, allowing in turn an easiest interpretation of the modeling. 3
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1002/cem.3097">doi:10.1002/cem.3097</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/s7trajomzbfchmbjlykxpaplpe">fatcat:s7trajomzbfchmbjlykxpaplpe</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20190503012546/https://hal.archives-ouvertes.fr/hal-01811514/file/sparseHDDA_v1.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/4d/54/4d5445e68e59d8f3979d3c8066ed84c289e33b83.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1002/cem.3097"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> wiley.com </button> </a>

not-MIWAE: Deep Generative Modelling with Missing not at Random Data [article]

Niels Bruun Ipsen, Pierre-Alexandre Mattei, Jes Frellsen
<span title="2021-03-18">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
This procedure has been enhanced by Mattei and Frellsen (2018a) using Metropolis-within-Gibbs.  ...  Similarly to Mattei and Frellsen (2019, equations (10,11) ), these moments can be estimated via self-normalised importance sampling.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2006.12871v2">arXiv:2006.12871v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/jsk4zkr45rgk3hz7mxtipf5ubi">fatcat:jsk4zkr45rgk3hz7mxtipf5ubi</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200625040250/https://arxiv.org/pdf/2006.12871v1.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> <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/a6/fe/a6fe8c1d5248cc778689cc3e4126abc4260439b6.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2006.12871v2" 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>

Don't fear the unlabelled: safe deep semi-supervised learning via simple debiasing [article]

Hugo Schmutz, Olivier Humbert, Pierre-Alexandre Mattei
<span title="2022-03-16">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Semi supervised learning (SSL) provides an effective means of leveraging unlabelled data to improve a model's performance. Even though the domain has received a considerable amount of attention in the past years, most methods present the common drawback of being unsafe. By safeness we mean the quality of not degrading a fully supervised model when including unlabelled data. Our starting point is to notice that the estimate of the risk that most discriminative SSL methods minimise is biased,
more &raquo; ... asymptotically. This bias makes these techniques untrustable without a proper validation set, but we propose a simple way of removing the bias. Our debiasing approach is straightforward to implement, and applicable to most deep SSL methods. We provide simple theoretical guarantees on the safeness of these modified methods, without having to rely on the strong assumptions on the data distribution that SSL theory usually requires. We evaluate debiased versions of different existing SSL methods and show that debiasing can compete with classic deep SSL techniques in various classic settings and even performs well when traditional SSL fails.
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2203.07512v2">arXiv:2203.07512v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/owyryibt2rh5pj4jvk2aknsi3i">fatcat:owyryibt2rh5pj4jvk2aknsi3i</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220318003228/https://arxiv.org/pdf/2203.07512v2.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/08/a7/08a7cd0d255f8bd449bdf7feae08091067133d65.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2203.07512v2" 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>

Localization and quantification of intramuscular damage using statistical parametric mapping and skeletal muscle parcellation

Alexandre Fouré, Arnaud Le Troter, Maxime Guye, Jean-Pierre Mattei, David Bendahan, Julien Gondin
<span title="2015-12-22">2015</span> <i title="Springer Nature"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/tnqhc2x2aneavcd3gx5h7mswhm" style="color: black;">Scientific Reports</a> </i> &nbsp;
In the present study, we proposed an original and robust methodology which combines the spatial normalization of skeletal muscle images, the statistical parametric mapping (SPM) analysis and the use of a specific parcellation in order to accurately localize and quantify the extent of skeletal muscle damage within the four heads of the quadriceps femoris. T 2 maps of thigh muscles were characterized before, two (D2) and four (D4) days after 40 maximal isometric electrically-evoked contractions
more &raquo; ... 25 healthy young males. On the basis of SPM analysis of coregistrated T 2 maps, the alterations were similarly detected at D2 and D4 in the superficial and distal regions of the vastus medialis (VM) whereas the proportion of altered muscle was higher in deep muscle regions of the vastus lateralis at D4 (deep: 35 ± 25%, superficial: 23 ± 15%) as compared to D2 (deep: 18 ± 13%, superficial: 17 ± 13%). The present methodology used for the first time on skeletal muscle would be of utmost interest to detect subtle intramuscular alterations not only for the diagnosis of muscular diseases but also for assessing the efficacy of potential therapeutic interventions and clinical treatment strategies. Unaccustomed exercises and neuromuscular diseases can lead to the occurrence of skeletal muscle damage 1-7 . These alterations are associated with several physiological events leading to inflammatory processes within muscle 1,6,8-11 . Magnetic resonance imaging (MRI) appears as a method of choice to investigate in vivo the extent of muscle damage in healthy subjects 12-16 or in patients with neuromuscular diseases 4,7,17,18 . MRI is a powerful non-invasive tool allowing for a spatially-resolved analysis of muscle tissue 4, 15, 19 . The increase in muscle proton transverse relaxation time (T 2 ) has been identified as a relevant biomarker of muscle damage 15,19-22 illustrating an inflammatory/edematous process 6,13,23-28 . Different T 2 changes among muscles have been already reported after exercise-induced muscle damage 15,21,22 and in dystrophic boys 4 . In most of these studies, the averaged T 2 value was calculated in regions of interest within a given muscle thereby ignoring any spatial information. Although local T 2 changes were assessed along muscles 19, 22 , no study provided information on the accurate localization and extent of intramuscular damage into the three dimensions (3D) of a skeletal muscle. Yet, accurate localization and quantification of muscle damage should provide more robust indices in diagnosis and longitudinal follow-ups of diseases or injuries. So far, only one study has reported information related to the distribution of intramuscular damage in dystrophic boys 7 . However, the analysis was only performed on a limited muscle volume (i.e., three MRI slices) and no information was provided on the localization of the most damaged areas within the muscle 7 . Moreover, we recently showed that neuromuscular electrostimulation (NMES) induced spatially heterogeneous T 2 changes in quadriceps femoris (QF) muscle group 22 with higher alterations in superficial muscles located beneath the stimulation electrodes (i.e., vastus lateralis [VL] and vastus medialis [VM]). On the sole basis of visual inspection of T 2 maps, we consistently observed that altered muscle areas were heterogeneously distributed within the damaged muscle. Due to large inter-individual morphologic differences (e.g., muscle/tendon length, muscle volume), the 3D coregistration of MR images appears essential to accurately
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1038/srep18580">doi:10.1038/srep18580</a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pubmed/26689827">pmid:26689827</a> <a target="_blank" rel="external noopener" href="https://pubmed.ncbi.nlm.nih.gov/PMC4686971/">pmcid:PMC4686971</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/bqdu7sfgyzg4lmhhx4lsxzglyy">fatcat:bqdu7sfgyzg4lmhhx4lsxzglyy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191207144639/http://europepmc.org/backend/ptpmcrender.fcgi?accid=PMC4686971&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/66/81/6681040b692ee26c08f5c6b29ddbd165413bb3c3.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1038/srep18580"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> Publisher / doi.org </button> </a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4686971" title="pubmed link"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> pubmed.gov </button> </a>
&laquo; Previous Showing results 1 &mdash; 15 out of 311 results