Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
481
datasets available to search
ShareScore release 0.9.0
Dataset results
481 results for “network model”
Dataset and scripts for manuscript "Using Neural Network Ensembles to Separate Ocean Biogeochemical and Physical Drivers of Phytoplankton Biogeography in Earth System Models"
<p>Please note: The title of this version contains an updated title for the manuscript compared to the previous version of this dataset. This is only due to title updates during the peer review process for the manuscript.</p> <p>The zip file contains the scripts, functions, and source files for the manuscript titled "Using Neural Network Ensembles to Separate Ocean Biogeochemical and Physical Drivers of Phytoplankton Biogeography in Earth System Models." The manuscript has been submitted for peer review.</p> <p>Please consult the README file for information on the specifications of the files.</p> <p>These files may occasionally be updated to add annotations to the scripts to make them more user friendly and to correct any errors.</p>
Supplementary files for Reconstructing Kinetic Models for Dynamical Studies of Metabolism using Generative Adversarial Networks, main part
<p><strong>Supplementary files containing datasets needed to reproduce the results of the manuscript "Reconstructing Kinetic Models for Dynamical Studies of Metabolism using Generative Adversarial Networks" by S. Choudhury et al.</strong></p> <p>The code to use with these data and reproduce the manuscript results is available at <a href="https://github.com/EPFL-LCSB/rekindle">https://github.com/EPFL-LCSB/rekindle</a> and <a href="https://gitlab.com/EPFL-LCSB/rekindle">https://gitlab.com/EPFL-LCSB/rekindle</a>. The execution of parts of this code is dependent on the SkimPy toolbox (<a href="https://github.com/EPFL-LCSB/skimpy">https://github.com/EPFL-LCSB/skimpy</a>). Refer to the readme files on the REKINDLE code repositories for more details.</p> <p><strong>Datasets:</strong></p> <ul> <li> <strong>models.zip </strong>- Datasets parameterizing kinetic nonlinear models of a wild-type <em>E. coli </em>strain used for training generative adversarial networks <ul> <li>subfolder 1: kinetic - contains the kinetic model (kin_varma_curated.yml)</li> <li>subfolder 2: thermo - contains the thermodynamic model for all the four physiologies (varma_fdp1, varma_fdp2, varma_fdp3, varma_fdp4)</li> <li>subfolder 3: steady_state_samples: contains the TFA steady state profiles for all four physiologies (samples_fdp1, sample_fdp2, samples_fdp3, samples_fdp4)</li> <li>subfolder 4: parameters - contains the kinetic parameter training dataset for each physiology (.hdf5 files), maximal eigenvalues (training labels) (maximal_eigenvalues.csv) and the minimum eigenvalues (minimal_eigenvalues.csv)</li> </ul> </li> <li><strong>vanilla_learning_training.zip:</strong> contains 4 folders for each of the 4 physiologies. <ul> <li>each of these folders contains 6 subsubfolders in the format N-<em>{n} </em>( N-10, N-50, N-100, N-500, N-1000, N-72000), where <em>{n} </em>represents the number of used training data samples.</li> <li>every subsubfolder N-{n} contains 5 repeats folders. Each repeat folder contains, <ul> <li>E_-1.npy - GAN generated kinetic parameters at E-th epoch/</li> <li>E_-1_max_eig.csv - the maximal eigenvalues of Jacobian for E_-1.npy (Note: eigenvalues were not calculated for N=10, 50, 100 as traning failed)/</li> </ul> </li> </ul> </li> <li><strong>transfer_learning_training.zip</strong> - contains 12 subfolders "tl_fdpi_fdpj" where i,j ={1,2,3,4} for each of the 12 transfer learning case <ul> <li>each of these folders contains 5 subsubfolders N-10, N-50, N-100, N-500, N-1000</li> <li>every subsubfolder N-{n} contains 5 repeats folders. Each repeat folder contains, <ul> <li>E_-1.npy - GAN generated kinetic parameters at E-th epoch/</li> <li>E_-1_max_eig.csv - the maximal eigenvalues of Jacobian for E_-1.npy </li> </ul> </li> </ul> </li> </ul> <ul> <li><strong>best_generators.zip</strong> <ul> <li>The best generators (with the highest incidence of relevant models) for each physiology (generator1- 4.h5)</li> <li>The normalizing scaling parameters for each generator (d_scaling.pkl).</li> <li> </li> </ul> </li> <li><strong>Temporal evolution of perturbations in non-linear ordinary differential equations</strong> <ul> <li><strong>vanilla_ODE_sample_parameters.zip</strong> - contains (i) 1000 REKINDLE generated kinetic parameter sets for each of the 4 physiologies and their corresponding eigenvalues (4 in total) (ii) 1000 ORACLE generated kinetic parameter sets for each of the 4 physiologies and their corresponding eigenvalues (4 in total). These parameter sets parameterize the ODEs which are integrated.</li> <li><strong>ode_solutions_physiology1.zip (available at </strong><a href="https://zenodo.org/record/5818192">https://zenodo.org/record/5818192</a><strong>) - </strong>contains 100 subfolders, each subfolder containing the time-series evolution data of 1000 kinetic models parameterized by REKINDLE generated parameter sets for physiology 1, each of the 1000 models having a random perturbation.</li> <li><strong>ode_solutions_physiology1_ORACLE.zip (available at </strong><a href="https://zenodo.org/record/5819669">https://zenodo.org/record/5819669</a><strong>) - </strong>contains 100 subfolders, each subfolder containing the time-series evolution data of 1000 kinetic models parameterized by ORACLE generated parameter sets for physiology 1, each of the 1000 models having a random perturbation.</li> <li><strong>ode_solutions_physiologies2-4.zip -</strong> contains 6 subfolders (physiology_2-4, physiology_2-4_ORACLE), with each subfolder containing 10 sub subfolders. Each sub subfolder containing the time-series evolution data of 1000 kinetic models parameterized by REKINDLE / ORACLE generated parameter sets for physiology 2-4, each of the 1000 models having a random perturbation.</li> <li><strong>transfer_learning_ODE_solutions.zip - </strong>contains two subfolders N_10, N_50, each subfolder contains 12 subsubfolders titled i_j (where i = {1,2,3,4} and j = {1,2,3,4} where 1_2 represent the transfer learning case from physiology 2 to physiology 1 and when using <em>{n}</em> samples from physiology 2 and so on (where <em>{n}</em>=10 and 50 respectively). Each subsubfolders contain <ul> <li>i_j.hdf5: contains 300 kinetic parameter sets generated using (i) REKINDLE for this transfer learning case</li> <li>i_j.csv: the maximal eigenvalues of the parameter sets</li> <li>solutions.csv: ODE integrated time series data for the relevant kinetic parameters out of the 300 generated.</li> </ul> </li> </ul> </li> </ul> <p> </p>
Supplementary files for Reconstructing Kinetic Models for Dynamical Studies of Metabolism using Generative Adversarial Networks, additional part 2
<p><strong>Supplementary files containing datasets needed to reproduce the results of the manuscript "Reconstructing Kinetic Models for Dynamical Studies of Metabolism using Generative Adversarial Networks" by S. Choudhury et al.</strong></p> <p>The code to use with these data and reproduce the manuscript results is available at <a href="https://github.com/EPFL-LCSB/rekindle">https://github.com/EPFL-LCSB/rekindle</a> and <a href="https://gitlab.com/EPFL-LCSB/rekindle">https://gitlab.com/EPFL-LCSB/rekindle</a>. The execution of parts of this code is dependent on the SkimPy toolbox (<a href="https://github.com/EPFL-LCSB/skimpy">https://github.com/EPFL-LCSB/skimpy</a>). Refer to the readme files on the REKINDLE code repositories for more details.</p> <ul> <li><strong>Temporal evolution of perturbations in non-linear ordinary differential equations</strong> <ul> <li><strong>ode_solutions_physiology1_ORACLE.zip</strong> - contains 100 subfolders, each subfolder containing the time-series evolution data of 1000 kinetic models parameterized by ORACLE generated parameter sets for physiology 1, each of the 1000 models having a random perturbation.</li> </ul> </li> </ul> <p>The detailed instructions and the main body of the dataset is available here: <a href="https://zenodo.org/record/5803120">https://zenodo.org/record/5803120</a></p>
Supplementary files for Reconstructing Kinetic Models for Dynamical Studies of Metabolism using Generative Adversarial Networks, additional part 1
<p><strong>Supplementary files containing datasets needed to reproduce the results of the manuscript "Reconstructing Kinetic Models for Dynamical Studies of Metabolism using Generative Adversarial Networks" by S. Choudhury et al.</strong></p> <p>The code to use with these data and reproduce the manuscript results is available at <a href="https://github.com/EPFL-LCSB/rekindle">https://github.com/EPFL-LCSB/rekindle</a> and <a href="https://gitlab.com/EPFL-LCSB/rekindle">https://gitlab.com/EPFL-LCSB/rekindle</a>. The execution of parts of this code is dependent on the SkimPy toolbox (<a href="https://github.com/EPFL-LCSB/skimpy">https://github.com/EPFL-LCSB/skimpy</a>). Refer to the readme files on the REKINDLE code repositories for more details.</p> <ul> <li><strong>Temporal evolution of perturbations in non-linear ordinary differential equations</strong> <ul> <li><strong>ode_solutions_physiology1.zip - </strong>contains 100 subfolders, each subfolder containing the time-series evolution data of 1000 kinetic models parameterized by REKINDLE generated parameter sets for physiology 1, each of the 1000 models having a random perturbation.</li> </ul> </li> </ul> <p>The detailed instructions and the main body of the dataset is available here: <a href="https://zenodo.org/record/5803120">https://zenodo.org/record/5803120</a></p>
Dataset: Tensor-network study of correlation-spreading dynamics in the two-dimensional Bose-Hubbard model
<p>Dataset</p> <p>Tensor-network study of correlation-spreading dynamics in the two-dimensional Bose-Hubbard model</p> <p>Ryui Kaneko, Ippei Danshita</p>
A tempοral Deep Convolutional Neural Network model on Sentinel-1 Image Time Series for pixel-wise Flood Classification (dataset)
<p>This is a dataset which has been designed to be used for flood time series classification. Each time series is annotated as flood or no-flood and represents a pixel-wise time series derived from stack of Sentinel-1 IW GRD images that have been pre-processed according to <a href="http://doi.org/10.5281/zenodo.6510223">https://doi.org/10.5281/zenodo.6510223</a>.</p>
Parallel generation of extensive vascular networks with application to an archetypal human kidney model
<p>Given the relevance of the inextricable coupling between microcirculation and physiology, and the relation to organ function and disease progression, the construction of synthetic vascular networks for mathematical modelling and computer simulation is becoming an increasingly broad field of research. Building vascular networks that mimic in-vivo morphometry is feasible through algorithms such as constrained constructive optimisation (CCO) and variations. Nevertheless, these methods are limited by the maximum number of vessels to be generated due to the whole network update required at each vessel addition. In this work, we propose a CCO-based approach endowed with a domain decomposition strategy to concurrently create vascular networks. The performance of this approach is evaluated by analysing the agreement with the sequentially generated networks and studying the scalability when building vascular networks up to 200,000 vascular segments. Finally, we apply our method to vascularise a highly complex geometry corresponding to the cortex of a prototypical human kidney. The technique presented in this work enables the automatic generation of extensive vascular networks, removing the limitation from previous works. Thus, we can extent vascular networks (e.g., obtained from medical images) to pre-arteriolar level, yielding patient-specific whole-organ vascular models with an unprecedented level of detail.</p>
Sparsified Model Zoo Twins: A Dataset of Sparsified Populations of Neural Network Models - MNIST
<p><strong>Abstract</strong></p> <p>In the last years, neural networks have evolved from laboratory environments to the state-of-the-art for many real-world problems. Our hypothesis is that neural network models (i.e., their weights and biases) evolve on unique, smooth trajectories in weight space during training. Following, a population of such neural network models (refereed to as “model zoo”) would form topological structures in weight space. We think that the geometry, curvature and smoothness of these structures contain information about the state of training and can be reveal latent properties of individual models. With such zoos, one could investigate novel approaches for (i) model analysis, (ii) discover unknown learning dynamics, (iii) learn rich representations of such populations, or (iv) exploit the model zoos for generative modelling of neural network weights and biases. Unfortunately, the lack of standardized model zoos and available benchmarks significantly increases the friction for further research about populations of neural networks. With this work, we publish a novel dataset of model zoos containing systematically generated and diverse populations of neural network models for further research. In total the proposed model zoo dataset is based on six image datasets, consist of 27 model zoos with varying hyperparameter combinations are generated and includes 50’360 unique neural network models resulting in over 2’585’360 collected model states. Additionally, to the model zoo data we provide an in-depth analysis of the zoos and provide benchmarks for multiple downstream tasks as mentioned before.</p> <p><strong>Dataset</strong></p> <p>This dataset is part of a larger collection of model zoos and contains the sparsified twins of models trained on MNIST. The original population is made available at https://doi.org/10.5281/zenodo.6632086. Sparsification is done using Variational Dropout, starting from the last epoch of the original population. The zip file contains the sparsification trajectory for 25 epochs for all 1000 models. All zoos with extensive information and code can be found at www.modelzoos.cc.</p> <p> </p>
Sparsified Model Zoo Twins: A Dataset of Sparsified Populations of Neural Network Models - SVHN
<p><strong>Abstract</strong></p> <p>In the last years, neural networks have evolved from laboratory environments to the state-of-the-art for many real-world problems. Our hypothesis is that neural network models (i.e., their weights and biases) evolve on unique, smooth trajectories in weight space during training. Following, a population of such neural network models (refereed to as “model zoo”) would form topological structures in weight space. We think that the geometry, curvature and smoothness of these structures contain information about the state of training and can be reveal latent properties of individual models. With such zoos, one could investigate novel approaches for (i) model analysis, (ii) discover unknown learning dynamics, (iii) learn rich representations of such populations, or (iv) exploit the model zoos for generative modelling of neural network weights and biases. Unfortunately, the lack of standardized model zoos and available benchmarks significantly increases the friction for further research about populations of neural networks. With this work, we publish a novel dataset of model zoos containing systematically generated and diverse populations of neural network models for further research. In total the proposed model zoo dataset is based on six image datasets, consist of 27 model zoos with varying hyperparameter combinations are generated and includes 50’360 unique neural network models resulting in over 2’585’360 collected model states. Additionally, to the model zoo data we provide an in-depth analysis of the zoos and provide benchmarks for multiple downstream tasks as mentioned before.</p> <p><strong>Dataset</strong></p> <p>This dataset is part of a larger collection of model zoos and contains the sparsified twins of models trained on SVHN. The original population is made available at https://doi.org/10.5281/zenodo.6632120. Sparsification is done using Variational Dropout, starting from the last epoch of the original population. The zip file contains the sparsification trajectory for 25 epochs for all 1000 models. All zoos with extensive information and code can be found at www.modelzoos.cc.</p>
WaterGEMS models for flow starved water transmission network
<p><strong>This data contains WaterGEMS models of a Water Transmission Network (WTN) in India. It deploys a newly developed modeling approach, Pressure Driven Analysis for Flow Straved Condition(PDA-FSC). It is used for evaluating the performance of flow-starved WTN. </strong></p>
Modelling the influence of parental effects on gene network evolution : Data and Program
<p>The C++ software "Simevolv" is a population genetics tool able to simulate the evolution of complex genetic architectures.</p> <p>This version is a development version from which the results of the manuscript "Modelling the influence of parental effects on gene network evolution" (in prep) have been obtained.</p> <p>The "Data" file corresponds to the simulation results used in the above mentioned article.</p>
Figure 3. Neural network representation for HMM given in Figure 2.-Neuroevolution Mechanism for Hidden Markov Model
<p>In our proposed structure, we injected a hidden layer to have a multilayer perceptron which<br> is more efficient than single layer perceptron.<br> To make this process clear, Figure 3 shows the neural networks for the HMM presented in<br> Figure 2.</p>
BRAIN Journal-Artificial Neuron Modelling Based on Wave Shape-Figure 1. Typical feedforward artificial neural network processing unit
<p>In Figure 1, each input X is weighted by a separate weight value w. These are then summed together to give a total input value for the neuron. This total can then be passed through a function, to transform it into the desired output value. This is then compared to the actual output value d; where the error or differences between the two sets is measured and used to correct the weight<br> values, to bring the two sets of values closer. One way to update the weights is after each individual pattern is presented and processed. Another option is to process the error after the presentation of the whole dataset, as part of a batch update.</p>
Dataset and Model Weights for Plasma Sheet Model Graph Network Simulator
<p>This repository contains the simulation data and pre-trained Graph Neural Network (GNN) models produced in [1].</p> <p>Two *.zip files are provided:</p> <ul> <li>data.zip - contains the datasets of train/test simulations produced using the Sheet Model algorithm [1, 2]</li> <li>models.zip - contains the GNN model weights (<em>*.</em>pkl<em>) </em>+ relevant training information and model parameters <em>(*.</em>yml<em> and *</em>.txt)</li> </ul> <p>Dataset subfolders are named according to dataset/{'train' or 'test'}/{number of sheets}/{boundary condition}/. Each subfolder contains multiple simulations and a single info.yml file with relevant information regarding the overall setup. For each i-th simulation the following files are provided:</p> <ul> <li>x_{i}.npy - array with sheet trajectories (#time-steps, #sheets)</li> <li>v_{i}.npy - array with sheet velocities (#time-steps, #sheets)</li> <li>x_eq_{i}.npy - array with sheet equilibrium positions (#time-steps, #sheets)</li> </ul> <p> Model sub-folders are named according to :</p> <ul> <li>models/{time step}/{seed} - default architecture (preferred)</li> <li>models/{time step}/{'collisions', 'nosent' or 'equivariant'}/{seed} - alternative (less performing) architectures mentioned in the paper appendices.</li> </ul> <p>For each model we provide:</p> <ul> <li>params_best.pkl - model weights that performed the best during training on the validation set</li> <li>params_final.pkl - model weights at the end of training</li> <li>model_cfg.yml - GNN architecture metadata</li> <li>train_cfg.yml - training configuration metadata</li> <li>train_data.yml - training dataset metadata</li> <li>loss.txt - training and validation loss per epoch</li> <li>loss_i.txt - training loss per gradient update step</li> </ul> <h3>Source Code</h3> <p>The source code used to produce the data, train, and test the models can be found at: <a href="https://github.com/diogodcarvalho/gns-sheet-model">https://github.com/diogodcarvalho/gns-sheet-model</a></p> <h3>References</h3> <p>[1] D. D. Carvalho, D. R. Ferreira, L. O. Silva, "Learning the dynamics of a one-dimensional plasma model with graph neural networks<em>", Mach. Learn.: Sci. Technol. 5 025048 </em>(2024)</p> <p>[2] J. Dawson, "One‐Dimensional Plasma Model"<em>, The Physics of Fluids</em> 5.4 (1962): 445-459.</p> <p> </p>
Physics-informed neural networks (PINNs) with unsaturated water flow models for inverse analysis of soil hydraulic parameters of layered soil profiles
<p>Information about the spatial distribution of soil hydraulic parameters is necessary for the accurate prediction of soil water flow and coupled movement of chemicals and heat at the field scale using a process-based model. Physics-informed neural networks (PINNs), which can provide physical constraints in deep learning to obtain a mesh-free solution, can be used to inversely estimate the soil hydraulic parameters from less and noisy training data. Previous studies using PINNs have successfully estimated soil hydraulic parameters for homogeneous soil but estimating such parameters of layered soil profiles where the interface depth and the parameters are unknown still has some difficulties. The objective of this study was to develop PINNs to inversely estimate the distribution of soil hydraulic parameters, such as saturated hydraulic conductivity and <em>α</em> and <em>n</em>, of the Mualem-van Genuchten model directly within layered soil profiles by predicting changes in pressure head from training data based on simulation results at given depths during infiltration. The impact of factors affecting PINNs performance, such as the weights assigned to each component of the loss function, the time range used in error computations, and the number of samples used to assess physical constraint was investigated. By assigning a larger weight to the physical constraint and excluding the earlier stage of infiltration in the loss function, the changes in pressure head and the three soil hydraulic parameter distributions within the layered soil profiles were successfully estimated. The developed PINNs can be further applied to more complex soils and can be improved.</p>
Code and Data for the Study "A User-Centric Model of Connectivity in Street Networks"
<p>This resource contains the code and results used in the paper:</p> <p>Corcoran, P. and R. Lewis (Pending) “A User-Centric Model of Connectivity in Street Networks”</p> <p>Please consult <strong>UserGuide.pdf</strong> for further information. </p>
Data & code repository for the article "A network toxicology approach for mechanistic modelling of nanomaterial hazard and adverse outcomes"
<p>This repository contains the relevant data and code supporting the study "A network toxicology approach for mechanistic modelling of nanomaterial hazard and adverse outcomes". </p> <p>In detail the following data sources have been included:</p> <ul> <li>the relevant code and supporting data (code_to_upload.zip and supporting_data.zip);</li> <li>supplementary materials of the paper, including: <ul> <li>individual enrichment results of the 93 exposures to the 31 ENMs (enrichments_results.zip);</li> <li>comparison between the mechanism of action retrieved from differentially expressed genes and network modelling (network_comparison_results.zip);</li> <li>overrepresented network edges in categories of networks (overrepresented_structures.zip)</li> </ul> </li> </ul>
An environmental resistance model to inform the biogeography of aquatic invasions in complex stream networks
<p>Freshwater invasions are a global conservation issue. Emerging tools for biogeographical analyses can provide critical information for their effective management and monitoring. Here, we propose a method to assess the distribution of environmental resistance of stream ecosystems to biological invasions by coupling multi‐stage habitat potential models for non‐native species. Location: Andean Patagonia (Chile and Argentina).Taxa: North American beaver (<em>Castor canadensis</em>), Chinook salmon (<em>Oncorhynchus tshawytscha</em>), and coho salmon (<em>O. kisutch</em>). Methods: Environmental resistance to invasive species was mapped throughout a large region of Patagonia by stacking multi‐stage habitat relationships for each target species and assessing the complementation between critical habitats at multiple scales. We generated an environmental model of stream networks derived from high‐resolution topographic and climatic data representing 15,406 drainage basins (>1 km2) covering an area of 369,791 km2. We quantified the intrinsic potential of stream reaches (100 m and 1000 m) to sustain high‐quality habitats and assessed habitat complementation (i.e., abundance and proximity) at the sub‐basin scale as a proxy for environmental resistance. Results: Our model revealed high heterogeneity in the distribution of environmental resistance to invasions throughout the study region, providing case‐specific insights for the research and management of invaders. Conclusions: Environmental resistance modelling is a novel method to study the biogeography of riverine invasions. Our approach is compatible with additional sources of information about species and the environment and shows versatility to diverse invasion scenarios and data sources. This method can be useful in prioritising research and management of incipient and spreading invasions, especially for large and data‐poor regions.</p>
Fig. 1 in Testing the robustness of transmission network models to predict ectoparasite loads. One lizard, two ticks and four years
Fig. 1. Transmission networks generated with (a) a short time window of infection; and (b) a long time window of infection, from the GPS location data of the lizards in the study population in 2010. Nodes represent individual lizards and edges between nodes are directed towards the lizard that is at risk of infection. The edges are weighted as described in the main text and the thicker the line the more weight is associated with that edge.
Supplementary material for publication "Multi-Echelon Inventory Optimization in Supply Chain Networks: Exploring Network Structures and Predictive Modeling"
<div> <div> <div> <p>This dataset collects different supply chain network structures generated artificially. We present four types of networks: Serial, Convergent, Divergent, and General, each type consisting of 20,000 individual instances. All 80,000 network instances generated are available to researchers and practitioners in Excel. The repository consists of separate files for each network instance consisting of each network inventory data, node connections, and a visual representation.</p> </div> </div> </div>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.