Filters








15,950 Hits in 6.1 sec

Optimal Single Sample Tests for Structured versus Unstructured Network Data [article]

Guy Bresler, Dheeraj Nagaraj
2018 arXiv   pre-print
We study the problem of testing, using only a single sample, between mean field distributions (like Curie-Weiss, Erdős-Rényi) and structured Gibbs distributions (like Ising model on sparse graphs and Exponential  ...  The test can distinguish the hypotheses with high probability above a certain threshold in the (inverse) temperature parameter, and is optimal in that below the threshold no test can distinguish the hypotheses  ...  In this paper we prove an abstract result, Theorem 3.4, that provides a framework for establishing near-optimal hypothesis tests between data from a network with a given dependency structure (like Ising  ... 
arXiv:1802.06186v2 fatcat:h3gvn6dejzgzjdqs56r444unxa

Personalized Federated Learning by Structured and Unstructured Pruning under Data Heterogeneity [article]

Saeed Vahidian and Mahdi Morafah and Bill Lin
2021 arXiv   pre-print
To realize this personalization, we leverage finding a small subnetwork for each client by applying hybrid pruning (combination of structured and unstructured pruning), and unstructured pruning.  ...  However, learning a single global model might not work well for all clients participating in the FL under data heterogeneity.  ...  Fig. 1 plots the test accuracy result of some sampled clients versus various pruning percentages when iteratively pruning by 5%-10% per iteration for CIFAR-10 on LeNet-5.  ... 
arXiv:2105.00562v2 fatcat:gag6ttfavjgy3j4xj6kp4hd3am

Strong, Tough, Stretchable, and Self-Adhesive Hydrogels from Intrinsically Unstructured Proteins

Mark A. Gonzalez, Joseph R. Simon, Ali Ghoorchian, Zachary Scholl, Shaoting Lin, Michael Rubinstein, Piotr Marszalek, Ashutosh Chilkoti, Gabriel P. López, Xuanhe Zhao
2017 Advanced Materials  
energy and maintain high elasticity of protein networks to enhance the strength and toughness of the resultant material. [13] Despite these successes, the synthesis of hydrogels containing highly structured  ...  Conversely, unstructured proteins (i.e., without tertiary structure) are relatively simple to design, can be easily synthesized with high yield, and can be biologically inert; [14] however, previously  ...  (Note: Error bars for all data are less than 1% and are smaller than the data point symbols.)  ... 
doi:10.1002/adma.201604743 pmid:28060425 fatcat:4l2ybmrzpfeudg6zthweit2xle

Natively Unstructured Loops Differ from Other Loops

Avner Schlessinger, Jinfeng Liu, Burkhard Rost
2007 PLoS Computational Biology  
In one application, NORSnet revealed previously undetected unstructured regions in putative targets for structural genomics and may thereby contribute to increasing structural coverage of large eukaryotic  ...  Training our new method, NORSnet, on predicted information rather than on experimental data yielded three major advantages: it removed the overlap between testing and training, it systematically covered  ...  Last, not least, thanks to all those who deposit their experimental data in public databases, and to those who maintain these databases, in particular to Keith Dunker and his colleagues for the maintenance  ... 
doi:10.1371/journal.pcbi.0030140 pmid:17658943 pmcid:PMC1924875 fatcat:zxnjdv65fbhl7cwtbv5nvp4qhe

Natively Unstructured Loops Differ from Other Loops

Avner Schlessinger, Jinfeng Liu, Burkhard Rost
2005 PLoS Computational Biology  
In one application, NORSnet revealed previously undetected unstructured regions in putative targets for structural genomics and may thereby contribute to increasing structural coverage of large eukaryotic  ...  Training our new method, NORSnet, on predicted information rather than on experimental data yielded three major advantages: it removed the overlap between testing and training, it systematically covered  ...  Last, not least, thanks to all those who deposit their experimental data in public databases, and to those who maintain these databases, in particular to Keith Dunker and his colleagues for the maintenance  ... 
doi:10.1371/journal.pcbi.0030140.eor fatcat:5r4bwlgegjfqplmc36oobheroy

Primer on machine learning

Parisa Rashidi, David A. Edwards, Patrick J. Tighe
2019 Current Opinion in Anaesthesiology  
This review provides a summary of key machine learning principles, as well as applications to both structured and unstructured datasets.  ...  Aside from increasing use in the analysis of electronic health record data, machine and deep learning algorithms are now key tools in the analyses of neuroimaging and facial expression recognition data  ...  Acknowledgements We gratefully acknowledge the assistance of Benjamin Shickel for his assistance in drafting Fig. 1 .  ... 
doi:10.1097/aco.0000000000000779 pmid:31408024 pmcid:PMC6785021 fatcat:7j7vvvfoezgzfotrvyks5irxne

A Framework for Neural Network Pruning Using Gibbs Distributions [article]

Alex Labach, Shahrokh Valaee
2020 arXiv   pre-print
This method can be used for either unstructured or structured pruning, and we provide explicit formulations for both.  ...  Neural network pruning is an important technique for creating efficient machine learning models that can run on edge devices.  ...  The authors would like to thank Fujitsu Laboratories Ltd. and Fujitsu Consulting (Canada) Inc. for providing financial support for this project at the University of Toronto.  ... 
arXiv:2006.04981v1 fatcat:4l4jwjelvrc7tgnynwhfbq332q

Predictive Structured-Unstructured Interactions in EHR Models: A Case Study of Suicide Prediction [article]

Ilkin Bayramli, Victor Castro, Yuval Barak-Corren, Emily M Madsen, Matthew K Nock, Jordan W Smoller, Ben Y Reis
2021 medRxiv   pre-print
Here, we compare the predictive value of structured and unstructured EHR data for predicting suicide risk.  ...  An NBC model trained on both structured and unstructured data yields similar performance (AUC = 0.743) to an NBC model trained on structured data alone (0.742, p = 0.668), while an RF model trained on  ...  Unstructured EHR data have been used for clinical predictive tasks, both as a standalone feature-set and in combination with structured data. 1 2 3 4 In order to optimally integrate both structured and  ... 
doi:10.1101/2021.08.10.21261831 fatcat:367umxbznnf6ratjjbqjxfvola

Multivariate time-series modeling with generative neural networks [article]

Marius Hofert, Avinash Prasad, Mu Zhu
2021 arXiv   pre-print
Following the popular copula-GARCH approach for modeling dependent MTS data, a framework based on a GMMN-GARCH approach is presented.  ...  Generative moment matching networks (GMMNs) are introduced as dependence models for the joint innovation distribution of multivariate time series (MTS).  ...  For both datasets, samples generated from the five GMMN models (see Section 4.1.2) more closely match the underlying cross-sectional dependence structure in their corresponding test datasets than those  ... 
arXiv:2002.10645v4 fatcat:slppflzs2javlbleoney45pdy4

Hierarchical Region-Network Sparsity for High-Dimensional Inference in Brain Imaging [chapter]

Danilo Bzdok, Michael Eickenberg, Gaël Varoquaux, Bertrand Thirion
2017 Lecture Notes in Computer Science  
Structured sparsity penalization has recently improved statistical models applied to high-dimensional data in various domains.  ...  Varying the relative importance of region and network structure within the hierarchical tree penalty captured complementary aspects of the neural activity patterns.  ...  Different emphasis on regions versus networks in hierarchical structured sparsity can yield very similar out-of-sample generalization.  ... 
doi:10.1007/978-3-319-59050-9_26 pmid:29743804 pmcid:PMC5937695 fatcat:dta6hxhsufbwbehgq5mtf5ehri

A scalable, efficient scheme for evaluation of stencil computations over unstructured meshes

James King, Robert M. Kirby
2013 Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis on - SC '13  
In this paper, we present an efficient method for performing stencil computations over unstructured meshes which increases data-locality and cache efficiency, and a scalable approach for stencil tiling  ...  Performing stencil operations over an unstructured mesh requires sampling of heterogeneous elements which often leads to inefficient memory access patterns and limits data locality/reuse.  ...  Structured meshes are easier to sample, allow for easy parallelization, and generally require no spatial data structure to manage.  ... 
doi:10.1145/2503210.2503214 dblp:conf/sc/0007K13 fatcat:nycucp7l2fbvtm2qbm53wkol5u

BoMuDANet: Unsupervised Adaptation for Visual Scene Understanding in Unstructured Driving Environments [article]

Divya Kothandaraman, Rohan Chandra, Dinesh Manocha
2021 arXiv   pre-print
A key aspect of our approach is that it can also identify objects that are encountered by the model for the fist time during the testing phase.  ...  We present an unsupervised adaptation approach for visual scene understanding in unstructured traffic environments.  ...  the network with adequate information for optimal performance on the complex target domain.  ... 
arXiv:2010.03523v3 fatcat:72sr7avhhjb7niwkheizlc2eeu

A Comparison of House Price Classification with Structured and Unstructured Text Data

Erika Cardenas, Connor Shorten, Taghi M. Khoshgoftaar, Borivoje Furht
2022 Proceedings of the ... International Florida Artificial Intelligence Research Society Conference  
Readers will gain an understanding of how to use machine learning models optimized with structured and unstructured text data to predict house prices.  ...  We report the performance of machine learning models trained with structured tabular representations and unstructured text descriptions.  ...  Tabular versus TF-IDF Representations of Houses We begin our experiments by comparing structured and unstructured data sources for house price classification.  ... 
doi:10.32473/flairs.v35i.130668 fatcat:oqwvqqun2zfjnaijiscy4rjajq

Deep Neural Networks for Multicomponent Molecular Systems

Kyohei Hanaoka
2020 ACS Omega  
However, their applicability remains limited to single-component materials and a general DNN model capable of handling various multicomponent molecular systems with composition data is still elusive, while  ...  Deep neural networks (DNNs) represent promising approaches to molecular machine learning (ML).  ...  sampling and 20 subsequent steps of optimization.  ... 
doi:10.1021/acsomega.0c02599 pmid:32875241 pmcid:PMC7450624 fatcat:eislnsigkjglrpkb23q5lhmwhu

Knowledge and Theme Discovery across Very Large Biological Data Sets Using Distributed Queries: A Prototype Combining Unstructured and Structured Data

Uma S. Mudunuri, Mohamad Khouja, Stephen Repetski, Girish Venkataraman, Anney Che, Brian T. Luke, F. Pascal Girard, Robert M. Stephens, Jan Aerts
2013 PLoS ONE  
As we move towards personalized medicine, the need to combine unstructured data, such as medical literature, with large amounts of highly structured and high-throughput data such as human variation or  ...  In order to achieve useful results, researchers require methods that consolidate, store and query combinations of structured and unstructured data sets efficiently and effectively.  ...  Acknowledgments The authors wish to thank Tom Plunkett for his technical contributions and edits to this manuscript, Ted Coyle for his ongoing Mahout/Hive counsel, Tyler Muth for his R expertise and detailed  ... 
doi:10.1371/journal.pone.0080503 pmid:24312478 pmcid:PMC3846626 fatcat:2or3whazpjc2lpg6wgz7qhp3ie
« Previous Showing results 1 — 15 out of 15,950 results