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.
921
datasets available to search
ShareScore release 0.7.1
Dataset results
921 results for “neural networks”
Project - Physics-informed neural network for lithium-ion battery degradation stable modeling and prognosis
<p>Here are the datasets for our publication entitled "<a href="https://www.nature.com/articles/s41467-024-48779-z">Physics-informed neural network for lithium-ion battery degradation stable modeling and prognosis</a>" published in Nature Communications. </p> <p>The object of this experiment is the 18650 nickel-cobalt-manganese (NCM) lithium-ion battery manufactured by "LISHEN". The chemical composition is LiNi<sub>0.5</sub>Co<sub>0.2</sub>Mn<sub>0.3</sub>O<sub>2</sub>. The nominal capacity of the battery is 2000 mAh, and the nominal voltage is 3.6 V. The charging cut-off voltage and discharging cut-off voltage are 4.2 V and 2.5 V, respectively. The whole experiment was conducted at room temperature. A total of 55 batteries were included in this experiment, conducted under 6 different charging and discharging strategies. The charging and discharging platform is ACTS-5V10A-GGS-D, and the sampling frequency for all data is 1Hz.</p> <p>Other details can be found in "Data Introduction.pdf" file.</p> <p>The <strong>Python Code</strong> for reading and preprocessing this dataset is available at: <a href="https://github.com/wang-fujin/Battery-dataset-preprocessing-code-library">https://github.com/wang-fujin/Battery-dataset-preprocessing-code-library</a></p> <p>Summary of articles using the this dataset: <a href="https://github.com/wang-fujin/XJTU-Battery-Dataset-Papers-Summary">https://github.com/wang-fujin/XJTU-Battery-Dataset-Papers-Summary</a></p> <p> </p> <p>If you find this data helpful, please consider citing our paper:</p> <p>Wang, F., Zhai, Z., Zhao, Z. <em>et al.</em> Physics-informed neural network for lithium-ion battery degradation stable modeling and prognosis. <em>Nat Commun</em> <strong>15</strong>, 4332 (2024). https://doi.org/10.1038/s41467-024-48779-z</p>
Code and partial data used in "Vertically recurrent neural networks for sub-grid parameterization"
<p>This repository contains the RNN training and evaluation code used in the paper<em> Vertically recurrent neural networks for sub-grid parameterization</em></p> <p> </p> <ul> <li> The radiative transfer emulation data can be accessed with through a Climetlab plugin (<a href="https://pypi.org/project/climetlab-maelstrom-radiation/">Climetlab-maelstrom-radiation</a>). <p>Datasets are downloaded and explained in the demo notebook here <a href="https://git.ecmwf.int/projects/MLFET/repos/maelstrom-radiation/browse/notebooks/demo_radiation.ipynb" rel="nofollow">https://git.ecmwf.int/projects/MLFET/repos/maelstrom-radiation/browse/notebooks/demo_radiation.ipynb</a></p> In addition the full training and testing code used in the paper is uploaded here (<em>pu-maelstrom-radiation.tar.gz</em>).</li> <li> </li> </ul> <p>Three parameterization problems from earlier studies are also included (we have modified the code from these papers to incorporate RNNs): </p> <ul> <li>non-orographic gravity wave drag (<a href="https://doi.org/10.1029/2021MS002477">Chantry et al. 2021</a>) <ul> <li>Based on TensorFlow</li> <li>This repository uses the <em>CliMetLab </em>plugin and<strong> downloads the data from the European Weather Cloud</strong></li> </ul> </li> <li>non-local parameterization (<a href="https://doi.org/10.1029/2022MS002984">Wang et al. 2022</a>) <ul> <li>The new code is based on TensorFlow, so you'll need both PyTorch and TensorFlow to run everything</li> <li><strong>See original paper for data access</strong></li> </ul> </li> <li>moist physics (Han et al. <a href="https://doi.org/10.1029/2022MS003508">2023</a>, <a href="https://doi.org/10.1029/2020MS002076">2020</a>) <ul> <li>Based on TensorFlow and PyTorch. This one has the most additions, e.g. code to generate a TensorFlow TFRecord dataset from the raw netCDF data archived in the original paper, autoregressive training and experimental model architectures in PyTorch</li> <li><strong>See original paper for data access</strong></li> </ul> </li> </ul> <p>Each of the code repos (unpack the tars) have an updated README.</p> <p>References:</p> <table> <tbody> <tr> <td> <div>Chantry, M., Hatfield, S., Dueben, P., Polichtchouk, I., & Palmer, T. (2021). Machine learning emulation of gravity wave drag in numerical weather forecasting. <em>Journal of Advances in Modeling Earth Systems</em>, <em>13</em>(7), e2021MS002477</div> <div> </div> <div> <div>Han, Y., Zhang, G. J., Huang, X., & Wang, Y. (2020). A moist physics parameterization based on deep learning. <em>Journal of Advances in Modeling Earth Systems</em>, <em>12</em>(9), e2020MS002076.</div> </div> <div> </div> <div>Han, Y., Zhang, G. J., & Wang, Y. (2023). An ensemble of neural networks for moist physics processes, its generalizability and stable integration. <em>Journal of Advances in Modeling Earth Systems</em>, <em>15</em>(10), e2022MS003508</div> <div> </div> <div>Wang, P., Yuval, J., & O’Gorman, P. A. (2022). Non‐local parameterization of atmospheric subgrid processes with neural networks. <em>Journal of Advances in Modeling Earth Systems</em>, <em>14</em>(10), e2022MS002984.</div> </td> </tr> <tr></tr> </tbody> </table> <div> </div>
Extra-P Version Used for Noise-Resilient Empirical Performance Modeling with Deep Neural Networks
<p>This is the Extra-P source code that was used for the analysis and evaluation of the IPDPS 2021 paper "Noise-Resilient Empirical Performance Modeling with Deep Neural Networks". It also contains the checkpoints and saved models for the DNN part of the adaptive modeler as well as the gathered synthetic evaluation data.</p>
Transformer-based graphical neural network with expert experience multimodal learning (TGEML) framework: a nanocomposite performance predictor
<p>TGEML is a novel multimodal nanocomposite processing framework consists of a polymer multimodal featurizer called TGEML-polymer and a nanoparticle expert experience featurizer called TGEML-nano.</p>
Video Supplement for Himes et al. (2024): "Using neural networks for near-real-time aerosol retrievals from OMPS Limb Profiler measurements"
<p>This archive contains the video supplement for</p> <p>Using neural networks for near-real-time aerosol retrievals from OMPS Limb Profiler measurements</p> <p>by Himes et al. (2024), submitted to Atmospheric Measurement Techniques. The file contains an animation of the V2.1 and NRT average retrieved extinction coefficient between 19.5--21.5 km at 997 nm for the 2024 Ruang eruptions.</p>
Forecasting of the Geomagnetic Activity for the Next 3 Days Utilizing Neural Networks Based on Parameters Related to Large-scale Structures of the Solar Corona
<p>These are supplementary data for the paper "Forecasting of the Geomagnetic Activity for the Next 3 Days Utilizing Neural Networks Based on Parameters Related to Large-scale Structures of the Solar Corona". They are:</p> <ul> <li>Python code to forecast Kp index</li> <li><span>Code to construct a nerual network model</span></li> </ul>
Dataset for Super-Resolution Image Reconstruction based on Random-coupled Neural Network and EDSR
<p>This dataset folder contains the DIV2K public dataset, which is utilized for model training and comprises 900 high-quality, high-resolution images along with their corresponding low-resolution versions. Additionally, all pre-trained models used in the experiment and their associated test results are publicly available.</p> <p>The main directory is organized into two subfolders: one labeled "dataset," which houses the DIV2K dataset, and another named "Model_results," which contains the pre-trained models and their corresponding test outcomes. The Dataset folder includes the original DIV2K dataset (referred to as "DIV2K") as well as a channel-expanded dataset processed by the RCNN model (designated as "DIV2K-RCNN"). Within the Model_results folder, the Model_trained subfolder contains all pre-trained models employed during the experiment, while the Test_results subfolder holds the test results for each model.</p>
Long short-term memory (LSTM) recurrent neural network for muscle activity detection
<p><strong>Background: </strong>The accurate temporal analysis of muscle activation is of great interest in many research areas, spanning<br> from neurorobotic systems to the assessment of altered locomotion patterns in orthopedic and neurological<br> patients and the monitoring of their motor rehabilitation. The performance of the existing muscle activity detectors<br> is strongly affected by both the SNR of the surface electromyography (sEMG) signals and the set of features used to<br> detect the activation intervals. This work aims at introducing and validating a powerful approach to detect muscle<br> activation intervals from sEMG signals, based on long short-term memory (LSTM) recurrent neural networks.<br> </p> <p><strong>Methods: </strong>First, the applicability of the proposed LSTM-based muscle activity detector (LSTM-MAD) is studied<br> through simulated sEMG signals, comparing the LSTM-MAD performance against other two widely used approaches,<br> i.e., the standard approach based on Teager–Kaiser Energy Operator (TKEO) and the traditional approach, used in<br> clinical gait analysis, based on a double-threshold statistical detector (Stat). Second, the effect of the Signal-to-Noise<br> Ratio (SNR) on the performance of the LSTM-MAD is assessed considering simulated signals with nine different SNR<br> values. Finally, the newly introduced approach is validated on real sEMG signals, acquired during both physiological<br> and pathological gait. Electromyography recordings from a total of 20 subjects (8 healthy individuals, 6 orthopedic<br> patients, and 6 neurological patients) were included in the analysis.</p> <p><strong>Results</strong>: The proposed algorithm overcomes the main limitations of the other tested approaches and it works<br> directly on sEMG signals, without the need for background-noise and SNR estimation (as in Stat). Results demonstrate<br> that LSTM-MAD outperforms the other approaches, revealing higher values of F1-score (F1-score > 0.91) and Jaccard<br> similarity index (Jaccard > 0.85), and lower values of onset/offset bias (average absolute bias < 6 ms), both on simulated<br> and real sEMG signals. Moreover, the advantages of using the LSTM-MAD algorithm are particularly evident for<br> signals featuring a low to medium SNR.</p> <p><strong>Conclusions</strong>: The presented approach LSTM-MAD revealed excellent performances against TKEO and Stat. The<br> validation carried out both on simulated and real signals, considering normal as well as pathological motor function<br> during locomotion, demonstrated that it can be considered a powerful tool in the accurate and effective recognition/<br> distinction of muscle activity from background noise in sEMG signals.</p> <p> </p>
Computer-aided Veress needle guidance using endoscopic optical coherence tomography and convolutional neural networks
<p>During laparoscopic surgery, the Veress needle is commonly used in pneumoperitoneum establishment. Precise placement of the Veress needle is still a challenge for the surgeon. In this study, a computer-aided endoscopic optical coherence tomography (OCT) system was developed to effectively and safely guide Veress needle insertion. This endoscopic system was tested by imaging subcutaneous fat, muscle, abdominal space, and the small intestine from swine samples to simulate the surgical process, including the situation with small intestine injury. Each tissue layer was visualized in OCT images with unique features and subsequently used to develop a system for automatic localization of the Veress needle tip by identifying tissue layers (or spaces) and estimating the needle-to-tissue distance. We used convolutional neural networks (CNNs) in automatic tissue classification and distance estimation. The average testing accuracy in tissue classification was 98.53±0.39%, and the average testing relative error in distance estimation reached 4.42±0.56% (36.09±4.92 μm).</p> <p>The dataset is split into two parts:<br> (1) <strong>Classification</strong>. The zip file <em>veress_classification_raw_images.zip</em> contains 40K images from 8 swine samples where there are 1K images per layer (skin, fat, muscle, abdominal space, and small intestine)<br> (2) <strong>Regression</strong>. The zip file <em>veress_regression_raw_images.zip</em><strong> </strong>contains 8K images of the abdominal space from the same 8 swine samples, and the ground truth distance labels for each sample are found in the Excel files <em>S[1-8]_distance_measurement_20210803.xlsx.</em></p>
Latent space images and related deep neural network for the BAGLS dataset
<p>In this repository, we provide the latent space images for the BAGLS (<a href="https://www.nature.com/articles/s41597-020-0526-3">Gómez, Kist et al., Sci Data 2020</a>, available at <a href="https://bagls.org/">www.bagls.org</a>) training dataset. We further provide a pre-trained deep neural network for glottis segmentation having only a single latent space, i.e. a latent space image, and no skip connections.</p>
Training data for benchtop NMR and UV/vis spectroscopy for Artificial Neural Networks
<p>Data set of low-field NMR spectra and UV/vis spectra for the synthesis of mesalazine intermediates, which were used as training or validation data for data processing with artificial neural networks development</p> <p><strong>Low-field NMR spectra for the nitration step:</strong></p> <p>The pure component spectrum of 2ClBA, 3N-2ClBA, and 5N-2ClBA are marked as NMR_pure_spectrum. The concentration levels for 2ClBA, 3N-2ClBA and 5N-2ClBA are in row 1, 2, and 3, respectively.</p> <p>The data sets marked as NMR_ represents low-field NMR-spectra recorded. The reference values for 2ClBA, 3N-2ClBA and 5N-2ClBA are in column 1, 2, and 3, respectively.</p> <p><strong>Datafusion data sets for the hydrolysis and nitration step</strong></p> <p>The NMR data are either recorded or simulated from the pure NMR spectrum of each individual component. The reference values for 2ClBA, 3N-2ClBA, 5N-2ClBA, 3-NSA and 5-NSA are either assigned with UHPLC measurements or calculated from the prepared solutions.</p> <p>The NMR spectra are depicted in datafusion_NMR_training. The reference values for 2ClBA, 3N-2ClBA and 5N-2ClBA are in column 1, 2, and 3, respectively.</p> <p>The UV/vis spectra are depicted in datafusion_UVvis_training. The reference values for 2ClBA, 3N-2ClBA, 5N-2ClBA, 3-NSA and 5-NSA are in column 1, 2, 3, 4, and 5, respectively.</p> <p><strong>Process data</strong></p> <p>The NMR spectra for the stability run and the run with dynamic changes are depicted in process_NMR_. The first column is the time stamp.</p> <p>The UV/vis spectra for the stability run and the run with dynamic changes are depicted in process_UV_. The first column is the time stamp.</p>
Convolutional neural network and data used for applied soundscape classification with Soundscapes 2 Landscapes (S2L)
<p>This repository documents the ABGQI-CNN manuscript (DOI: <a href="https://doi.org/10.1016/j.ecolind.2022.108831">https://doi.org/10.1016/j.ecolind.2022.108831</a>). It contains supplementary materials, data used to train a soundscape classification convolutional neural network (CNN), and data to generate manuscript results. The accompanying code can be found at <a href="https://doi.org/10.5281/zenodo.6038460">https://doi.org/10.5281/zenodo.6038459</a>. Files include:</p> <ul> <li><strong>ABGQI-CNN.tar: </strong>saved CNN model weights for the 5-class soundscape classifier using a MobileNetV2 architecture pre-trained with bird vocalization data.</li> <li><strong>ABGQI_mel_spectrograms.tar</strong>: spectrograms used for fine-tuning the pre-trained CNN, above, with training, validation, and testing data splits.</li> <li><strong>freesound_licensing.csv</strong>: file names and license information related to Freesound auxiliary files.</li> <li><strong>RavenLite_Training_Data_Collection.pdf</strong>: a manual for RavenLite ROI annotation.</li> <li><strong>S2L_site_geog-env_data.csv</strong>: environmental and geographic data (sans GPS) related to site locations in S2L project 2017-2020.</li> <li><strong>site_avg_ABGQIU_fscore_075_daytime.csv</strong>: the average site rate of soundscape components for 5 a.m. to 8 p.m.</li> <li><strong>site_by_hour_ABGQIU_fscore_075.csv</strong>: the average hourly site rate of soundscape components</li> <li><strong>site_classifications_beta075.tar</strong>: a directory containing a CSV for every site with threshold optimized classifications for each 2-s Mel spectrogram</li> <li><strong>site_prediction_probabilies.tar</strong>: a directory containing a CSV for every site with ABGQI-CNN probabilities for each 2-s Mel spectrogram</li> <li><strong>Supplementary_Materials.pdf</strong>: includes additional material and analyses related to the accompanying manuscript. </li> </ul> <p>Contact Colin Quinn at cq73@nau.edu for questions related to this repository or if you have an interest in the original wav recordings. Please be aware that underlying software, specifically for the CNN implementation, may not continue stability as python libraries are updated.</p>
4D-Var data assimilation experiment of the Lorenz 96 model using an adjoint model of a neural network surrogate model
<p>These data are the output of the 4D-Var data assimilation experiment of the Lorenz96 model using an adjoint model of a neural network surrogate model.<br> The details are described in Nishizawa (2022).<br> </p>
Data sets used in "Neural network processing of holographic images"
<p>Included are the training, validation, and testing data sets for synthetic holograms (netCDF), the HOLODEC data set containing the RF07 examples (netCDF), and the two splits of manually labeled HOLODEC image tiles (numpy arrays). The source code for using the data sets can be found at https://github.com/NCAR/holodec-ml </p>
Data from: Applicability of artificial neural networks to integrate socio-technical drivers of buildings recovery following extreme wind events
<p>The data provided and the associated MATLAB code were used to build an Artificial Neural Network Model to capture the reconstruction (recovery) of various buildings subjected to tornado events in the State of Missouri. The ANN model utilizes relevant tornado, societal demographic, and structural data to determine a building's resulting damage state from an extreme wind event and the subsequent recovery time. Abstract for the publication is as follows:</p> <p>In a companion article, previously published in Royal Society Open Science, the authors used Graph Theory to evaluate artificial neural network models for potential social and building variables interactions contributing to building wind damage. The results promisingly highlighted the importance of social variables in modeling damage as opposed to the traditional approach of solely considering physical characteristics of a building. Within this update article, the same methods are used to evaluate two different artificial neural networks for modelling building repair and/or rebuild (recovery) time. In contrast to the damage models, the recovery models consider (A) primarily social variables and then (B) introduce structural variables. These two models are then evaluated using centrality and shortest path concepts of Graph Theory as well as validated against data from the 2011 Joplin Tornado. The results of this analysis do not show the same distinctions as were found in the analysis of the damage models from the companion article. The overarching lack of discernible and consistent differences in the recovery models suggests that social variables that drive damage are not necessarily contributions to recovery. The differences also serve to reinforce that machine learning methods are best used when the contributing variables are already well understood.</p>
Fused Image dataset for convolutional neural Network-based crack Detection (FIND)
<p>The “<strong>F</strong>used <strong>I</strong>mage dataset for convolutional neural <strong>N</strong>etwork-based crack <strong>D</strong>etection” (<strong>FIND</strong>) is a large-scale image dataset with pixel-level ground truth crack data for deep learning-based crack segmentation analysis. It features four types of image data including raw intensity image, raw range (i.e., elevation) image, filtered range image, and fused raw image. The FIND dataset consists of 2500 image patches (dimension: 256x256 pixels) and their ground truth crack maps for each of the four data types.</p> <p>The images contained in this dataset were collected from multiple bridge decks and roadways under real-world conditions. A laser scanning device was adopted for data acquisition such that the captured raw intensity and raw range images have pixel-to-pixel location correspondence (i.e., spatial co-registration feature). The filtered range data were generated by applying frequency domain filtering to eliminate image disturbances (e.g., surface variations, and grooved patterns) from the raw range data [1]. The fused image data were obtained by combining the raw range and raw intensity data to achieve cross-domain feature correlation [2,3]. Please refer to [4] for a comprehensive benchmark study performed using the FIND dataset to investigate the impact from different types of image data on deep convolutional neural network (DCNN) performance.</p> <p>If you share or use this dataset, please cite [4] and [5] in any relevant documentation. </p> <p>In addition, an image dataset for crack classification has also been published at [6].</p> <p>References:</p> <p>[1] Shanglian Zhou, & Wei Song. (2020). Robust Image-Based Surface Crack Detection Using Range Data. Journal of Computing in Civil Engineering, 34(2), 04019054. <a href="https://doi.org/10.1061/(asce)cp.1943-5487.0000873">https://doi.org/10.1061/(asce)cp.1943-5487.0000873</a></p> <p>[2] Shanglian Zhou, & Wei Song. (2021). Crack segmentation through deep convolutional neural networks and heterogeneous image fusion. Automation in Construction, 125. <a href="https://doi.org/10.1016/j.autcon.2021.103605">https://doi.org/10.1016/j.autcon.2021.103605</a></p> <p>[3] Shanglian Zhou, & Wei Song. (2020). Deep learning–based roadway crack classification with heterogeneous image data fusion. Structural Health Monitoring, 20(3), 1274-1293. <a href="https://doi.org/10.1177/1475921720948434">https://doi.org/10.1177/1475921720948434</a> </p> <p>[4] Shanglian Zhou, Carlos Canchila, & Wei Song. (2023). Deep learning-based crack segmentation for civil infrastructure: data types, architectures, and benchmarked performance. Automation in Construction, 146. <a href="https://doi.org/10.1016/j.autcon.2022.104678">https://doi.org/10.1016/j.autcon.2022.104678</a></p> <p>[5] (<strong>This dataset</strong>) Shanglian Zhou, Carlos Canchila, & Wei Song. (2022). Fused Image dataset for convolutional neural Network-based crack Detection (FIND) [Data set]. Zenodo. <a href="https://doi.org/10.5281/zenodo.6383044">https://doi.org/10.5281/zenodo.6383044</a></p> <p>[6] Wei Song, & Shanglian Zhou. (2020). Laser-scanned roadway range image dataset (LRRD). Laser-scanned Range Image Dataset from Asphalt and Concrete Roadways for DCNN-based Crack Classification, DesignSafe-CI. <a href="https://doi.org/10.17603/ds2-bzv3-nc78">https://doi.org/10.17603/ds2-bzv3-nc78</a></p>
Protein vibrational frequencies dataset: Rapid Prediction of Protein Natural Frequencies using Graph Neural Networks
<p>Dataset for machine learning model, based on graph neural network, to predict protein natural frequencies using Graph Neural Networks. </p> <p><strong>Code</strong>: https://github.com/lamm-mit/ProteinMechanicsGNN</p> <p><strong>Paper</strong>: </p> <p>Rapid Prediction of Protein Natural Frequencies using Graph Neural Networks</p> <p>Kai Guo and Markus J. Buehler</p> <p><em>Digital Discovery</em>, 2022, DOI: 10.1039/D1DD00007A</p>
Understanding the Influence of Receptive Field and Network Complexity in Neural-Network-Guided TEM Image Analysis
<p>TEM images of Au nanoparticles of various sizes on ultra-thin carbon substrates and their corresponding labels for semantic segmentation. The images have a dataset label of "images" and the labels have a dataset label of "labels". </p>
Echolocation clicks and anthropogenic detections with neural network labels in Hawaiian Island HARP data from Kona, Kaua`i, and Pearl and Hermes Reef
<p><span>This dataset consists of echolocation clicks and detections of anthropogenic signals at three sites in the Hawaiian Islands Archipelago. These sites are </span><span>Hawaii/Hawaii_K, </span><span>Kauai/KA, and </span><span>Pearl and Hermes Reef/PHR. </span><span>Echolocation clicks were grouped into 5 minute bins, for which summary data is provided. Files are in .mat format that can be read using any desired coding language using a netcdf reading script. Files are separated by site, deployment, and neural network class (i.e. sitedeployment_cbins_class or site_deployment_cbins_class). Manual labels are provided.</span></p>
A Tungsten Deep Neural-Network Potential for Simulating Mechanical Property Degradation Under Fusion Service Environment
<p>The DP-HYB and DP-SE2potential and the W training database.</p>
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.