Filters








12 Hits in 1.4 sec

Balsam: Automated Scheduling and Execution of Dynamic, Data-Intensive HPC Workflows [article]

Michael A. Salim, Thomas D. Uram, J. Taylor Childers, Prasanna Balaprakash, Venkatram Vishwanath, Michael E. Papka
<span title="2019-09-18">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
The efficacy of this system is illustrated using two case studies: hyperparameter optimization of deep neural networks, and high-throughput single-point quantum chemistry calculations.  ...  With abstractions for the local resource scheduler and MPI environment, Balsam dynamically packages tasks into ensemble jobs and manages their scheduling lifecycle.  ...  DeepHyper We interfaced Balsam with the DeepHyper [2] package for large-scale hyperparameter searches on deep neural networks.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1909.08704v1">arXiv:1909.08704v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/wqfpcf2vozbdzejvzqh3qrc75i">fatcat:wqfpcf2vozbdzejvzqh3qrc75i</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200902063934/https://arxiv.org/pdf/1909.08704v1.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/2b/ba/2bbac36cc3bae1474d85fe5c2f8de4eefd90efd4.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1909.08704v1" 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-Ensemble-Based Uncertainty Quantification in Spatiotemporal Graph Neural Networks for Traffic Forecasting [article]

Tanwi Mallick, Prasanna Balaprakash, Jane Macfarlane
<span title="2022-04-05">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We focus on a diffusion convolutional recurrent neural network (DCRNN), a state-of-the-art method for short-term traffic forecasting.  ...  We develop a scalable deep ensemble approach to quantify uncertainties for DCRNN.  ...  Given the training and validation data, a neural network model, and the feasible set of values of the training hyperparameters, DeepHyper uses an asynchronous BO method seeks to minimize the validation  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2204.01618v2">arXiv:2204.01618v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/rmd7j2d5dbhafevr7b2qhhw2mm">fatcat:rmd7j2d5dbhafevr7b2qhhw2mm</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220406205939/https://arxiv.org/pdf/2204.01618v1.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/47/f2/47f24976434c81618b2816fbe7830cee9ee32bce.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2204.01618v2" 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>

HYPPO: A Surrogate-Based Multi-Level Parallelism Tool for Hyperparameter Optimization [article]

Vincent Dumont, Casey Garner, Anuradha Trivedi, Chelsea Jones, Vidya Ganapati, Juliane Mueller, Talita Perciano, Mariam Kiran, Marc Day
<span title="2021-10-04">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We present a new software, HYPPO, that enables the automatic tuning of hyperparameters of various deep learning (DL) models.  ...  Unlike other hyperparameter optimization (HPO) methods, HYPPO uses adaptive surrogate models and directly accounts for uncertainty in model predictions to find accurate and reliable models that make robust  ...  The main contributions of HYPPO include (1) an automated and adaptive search for hyperparameters using uncertainty quantification to evolve the best deep neural network solutions for both PyTorch and Tensorflow  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2110.01698v1">arXiv:2110.01698v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/xscwlvuo5bf5njyywgwx4aq5iy">fatcat:xscwlvuo5bf5njyywgwx4aq5iy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20211007095307/https://arxiv.org/pdf/2110.01698v1.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/c8/76/c8763f16e4dbad165cc5b4fde37f03b1df0af772.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2110.01698v1" 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>

AgEBO-Tabular: Joint Neural Architecture and Hyperparameter Search with Autotuned Data-Parallel Training for Tabular Data [article]

Romain Egele, Prasanna Balaprakash, Venkatram Vishwanath, Isabelle Guyon, Zhengying Liu
<span title="2021-10-26">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Neural architecture search (NAS) is an AutoML approach that generates and evaluates multiple neural network architectures concurrently and improves the accuracy of the generated models iteratively.  ...  method for tuning the hyperparameters of the data-parallel training simultaneously.  ...  for joint neural architecture and hyperparameter search.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2010.16358v2">arXiv:2010.16358v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/4sajee34xrfhfblx4rmoantiby">fatcat:4sajee34xrfhfblx4rmoantiby</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20211113052139/https://arxiv.org/pdf/2010.16358v2.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/b5/8a/b58a229c39f465d23a9e67e4de00667c52d357a8.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2010.16358v2" 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>

Recurrent Neural Network Architecture Search for Geophysical Emulation [article]

Romit Maulik, Romain Egele, Bethany Lusch, Prasanna Balaprakash
<span title="2020-08-13">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Constructing neural networks for forecasting such data is nontrivial, however, and often requires trial and error.  ...  We develop a scalable neural architecture search for generating stacked LSTMs to forecast temperature in the NOAA Optimum Interpolation Sea-Surface Temperature data set.  ...  DeepHyper is a scalable open-source hyperparameter and NAS package that was previously assessed for automated discovery of fully connected neural networks on tabular data.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2004.10928v2">arXiv:2004.10928v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/2p5wmsuu4jcopl6qohnbs4jbma">fatcat:2p5wmsuu4jcopl6qohnbs4jbma</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200819041050/https://arxiv.org/pdf/2004.10928v2.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/2004.10928v2" 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>

Neuromorphic Architecture Optimization for Task-Specific Dynamic Learning

Sandeep Madireddy, Angel Yanguas-Gil, Prasanna Balaprakash
<span title="">2019</span> <i title="ACM Press"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/fu5ttfvpovd5vm2zyen3mypunm" style="color: black;">Proceedings of the International Conference on Neuromorphic Systems - ICONS &#39;19</a> </i> &nbsp;
We employ DeepHyper, a scalable, asynchronous model-based search, to simultaneously optimize the choice of meta-learning rules and their hyperparameters.  ...  A key challenge, however, is to understand which learning rules are best suited for specific tasks and how the relevant hyperparameters can be fine-tuned.  ...  We then employ a scalable, asynchronous model-based search (AMBS) algorithm to simultaneously optimize learning rules and their hyperparameters.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1145/3354265.3354270">doi:10.1145/3354265.3354270</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/icons2/MadireddyYB19.html">dblp:conf/icons2/MadireddyYB19</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/nt5dblse5nbalmu3rprtlb5pya">fatcat:nt5dblse5nbalmu3rprtlb5pya</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200826185856/https://arxiv.org/pdf/1906.01668v1.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/4a/84/4a84c7534907ee0f7bc3b5d29c1a113cb490b563.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1145/3354265.3354270"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> acm.org </button> </a>

Scalable reinforcement-learning-based neural architecture search for cancer deep learning research

Prasanna Balaprakash, Romain Egele, Misha Salim, Stefan Wild, Venkatram Vishwanath, Fangfang Xia, Tom Brettin, Rick Stevens
<span title="">2019</span> <i title="ACM Press"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/zigbcra6rjdivda6lkzknwuo5q" style="color: black;">Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis on - SC &#39;19</a> </i> &nbsp;
To that end, we develop a reinforcement-learning-based neural architecture search to automate deep-learning-based predictive model development for a class of representative cancer data.  ...  We show that our approach discovers deep neural network architectures that have significantly fewer trainable parameters, shorter training time, and accuracy similar to or higher than those of manually  ...  Hyperparameter search approaches try to find the best values for the hyperparameters for a fixed neural architecture.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1145/3295500.3356202">doi:10.1145/3295500.3356202</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/sc/BalaprakashESWV19.html">dblp:conf/sc/BalaprakashESWV19</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/t3pbc6oruffgraj7tiedv4ekoi">fatcat:t3pbc6oruffgraj7tiedv4ekoi</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200912164922/https://arxiv.org/pdf/1909.00311v1.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/24/45/24454f77f76a41c1bf63899e79240e367d3cf648.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1145/3295500.3356202"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> acm.org </button> </a>

Data-Driven Modeling of Coarse Mesh Turbulence for Reactor Transient Analysis Using Convolutional Recurrent Neural Networks [article]

Yang Liu, Rui Hu, Adam Kraus, Prasanna Balaprakash, Aleksandr Obabko
<span title="2021-11-06">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
A novel neural network architecture, combining a densely connected convolutional network and a long-short-term-memory network, is developed that can efficiently learn from the spatial-temporal CFD transient  ...  The neural network model was trained and optimized on a loss-of-flow transient and demonstrated high accuracy in predicting the turbulent viscosity field during the whole transient.  ...  Following the Bayesian optimization framework, DeepHyper uses an asynchronous model-based search to obtain an optimized set of hyperparameters with minimal RMSE on the testing dataset.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2109.04423v2">arXiv:2109.04423v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/kkwi57vfnneypempm4mwpkivvq">fatcat:kkwi57vfnneypempm4mwpkivvq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20211110001919/https://arxiv.org/pdf/2109.04423v2.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/cf/54/cf542593f39467f562ab4251922120d629f2071f.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2109.04423v2" 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>

Non-autoregressive time-series methods for stable parametric reduced-order models [article]

Romit Maulik, Bethany Lusch, Prasanna Balaprakash
<span title="2020-06-25">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
To address these issues, we have developed a non-autoregressive time series approach for predicting linear reduced-basis time histories of forward models.  ...  There has been significant interest in the use of techniques borrowed from machine learning to reduce the computational expense and/or improve the accuracy of predictions for these systems.  ...  For optimizing the hyperparameters of the different methods, we use DeepHyper [38] , a hyperparameter search package that leverages Bayesian black-box optimization from scikit-optimize [39] .  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2006.14725v1">arXiv:2006.14725v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/3emv6zydzjh75c6b4pbwb6crfq">fatcat:3emv6zydzjh75c6b4pbwb6crfq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200630004315/https://arxiv.org/pdf/2006.14725v1.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/cb/33/cb33ac00fac9ac92a38083a75759016dc3fe61ce.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2006.14725v1" 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 data-centric weak supervised learning for highway traffic incident detection [article]

Yixuan Sun, Tanwi Mallick, Prasanna Balaprakash, Jane Macfarlane
<span title="2021-12-17">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
setup for final detection.  ...  Using the data from loop detector sensors for near-real-time detection of traffic incidents in highways is crucial to averting major traffic congestion.  ...  DeepHyper: Asynchronous hyperparameter search for deep neural networks. In: 2018 IEEE 25th international conference on High Performance Computing (HiPC). IEEE; 2018. p. 42–51. [11] Xiao J.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2112.09792v1">arXiv:2112.09792v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/chvrvdeiardjzcojeppf3e4bce">fatcat:chvrvdeiardjzcojeppf3e4bce</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20211231105107/https://arxiv.org/pdf/2112.09792v1.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/4a/8f/4a8f82579e1d9529e496d34efd58fbb6152c2ac1.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2112.09792v1" 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>

Colmena: Scalable Machine-Learning-Based Steering of Ensemble Simulations for High Performance Computing [article]

Logan Ward, Ganesh Sivaraman, J. Gregory Pauloski, Yadu Babuji, Ryan Chard, Naveen Dandu, Paul C. Redfern, Rajeev S. Assary, Kyle Chard, Larry A. Curtiss, Rajeev Thakur, Ian Foster
<span title="2021-10-06">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
by a factor of 100 over unguided searches.  ...  Methods that use machine learning (ML) to create proxy models of simulations show particular promise for guiding ensembles but are challenging to deploy because of the need to coordinate dynamic mixes  ...  Tools that perform hyperparameter searches for neural network design, such as DeepHyper [12] and Tune [39] , illustrate how to handle ML tasks at scale.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2110.02827v1">arXiv:2110.02827v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/fp5lul5j3bcivjzy5qfvwxzxoe">fatcat:fp5lul5j3bcivjzy5qfvwxzxoe</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20211009215251/https://arxiv.org/pdf/2110.02827v1.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/9f/65/9f65f4d242ff6a63c57c869df13cc20418b3ab8f.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2110.02827v1" 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>

Three dimensions, two microscopes, one code: Automatic differentiation for x-ray nanotomography beyond the depth of focus limit

Ming Du, Youssef S. G. Nashed, Saugat Kandel, Doğa Gürsoy, Chris Jacobsen
<span title="2020-03-27">2020</span> <i title="American Association for the Advancement of Science (AAAS)"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/rcuo52vymzdnhaszizbal7vbki" style="color: black;">Science Advances</a> </i> &nbsp;
We show that the same proposed method works for both full-field microscopy and for coherent scanning techniques like ptychography.  ...  Our implementation uses the optimization toolbox and the automatic differentiation capability of the open-source deep learning package TensorFlow, demonstrating a straightforward way to solve optimization  ...  While one can use a progressive strategy (optimize one parameter with others fixed and then move to the next one) to search the hyperparameter space, there are also dedicated packages for parameter searching  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1126/sciadv.aay3700">doi:10.1126/sciadv.aay3700</a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pubmed/32258397">pmid:32258397</a> <a target="_blank" rel="external noopener" href="https://pubmed.ncbi.nlm.nih.gov/PMC7101216/">pmcid:PMC7101216</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ggc7ypaec5cdtgds5l77pvqnyy">fatcat:ggc7ypaec5cdtgds5l77pvqnyy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200327192940/https://advances.sciencemag.org/content/advances/6/13/eaay3700.full.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 noreferrer" href="https://doi.org/10.1126/sciadv.aay3700"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> sciencemag.org </button> </a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7101216" title="pubmed link"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> pubmed.gov </button> </a>