Find research datasets worth reusing

opencc-by-4.0Dec 2022View details →

Dataset of "Near-real-time diagnosis of electron optical phase aberrations in scanning transmission electron microscopy using an artificial neural network"

Dataset containing the jupyter notebook used to construct the database of image, to model and train ANN and to analyze the experimental data. Furthermore there are also a reduced database of 100 images that can be utilized to test the ANN, the h5 file containing the ANN weigths and other supporting files.

opencc-by-4.0Apr 2023View details →

NsCircle datasets from "Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics"

Datasets with the simulations of the incompressible flow around an elliptical as described by the incompressible Navier-Stokes equations. These simulations were used to train and test the MuS-GNN models in the paper:     Multi-scale rotation-equivariant graph neural networks for     unsteady Eulerian fluid dynamics (https://doi.org/10.1063/5.0097679) The datasets are:   - train/NsEllipse   - test/NsEllipseLowRe   - test/NsEllipseHighRe   - test/NsEllipseThin   - test/NsEllipseThick   - test/NsEllipseNarrow   - test/NsEllipseWide   - test/NsEllipseAoA     To cite these datasets, use the following reference: Mario Lino, Stathi Fotiadis, Anil A. Bharath, and Chris Cantwell. "Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics". Physics of Fluids, 34 (2022). @article{lino2022multi,     author = {Lino, Mario and Fotiadis, Stathi and Bharath, Anil A. and Cantwell, Chris},     title = {{Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics}},     journal = {Physics of Fluids},     volume = {34},     year = {2022},     url = {https://doi.org/10.1063/5.0097679}, }

opencc-by-4.0Apr 2023View details →

Uncovering local aggregated air quality index with smartphone captured images leveraging efficient deep convolutional neural network

Short Description: In this research, we vigorously analyze the difficulties of predicting location-specific PM2.5 concentration from photos captured by smartphone cameras. Here, we particularly focus on Dhaka, the capital of Bangladesh, considering its very high level of air pollution exposure to a huge number of its dwellers. In our research, we develop a Deep Convolutional Neural Network (DCNN) and train it using more than a thousand outdoor photos captured and labeled by us. We capture the photos at various locations in Dhaka, Bangladesh, and label them based on PM2.5 concentration data extracted from the local US consulate as computed by the NowCast algorithm. During training with the dataset, our model learns a correlation index through supervised learning, which improves the model's ability to act as a Picture-based Predictor of PM2.5 Concentration (PPPC) making it capable of detecting comparable daily aggregated AQI index from a photo captured by a smartphone. Code and More Details: https://github.com/lepotatoguy/aqi

NsEllipse datasets from "Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics"

Datasets with simulations of the incompressible flow around an elliptical cylinder as described by the incompressible Navier-Stokes equations. These simulations were used to train and test the MuS-GNN models in the paper:     "Multi-scale rotation-equivariant graph neural networks for     unsteady Eulerian fluid dynamics" (https://doi.org/10.1063/5.0097679) The datasets are:   - train/NsEllipse   - test/NsEllipseLowRe   - test/NsEllipseHighRe   - test/NsEllipseThin   - test/NsEllipseThick   - test/NsEllipseNarrow   - test/NsEllipseWide   - test/NsEllipseAoA   To cite these datasets, use the following reference: Mario Lino, Stathi Fotiadis, Anil A. Bharath, and Chris Cantwell. "Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics". Physics of Fluids, 34 (2022). @article{lino2022multi,     author = {Lino, Mario and Fotiadis, Stathi and Bharath, Anil A. and Cantwell, Chris},     title = {{Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics}},     journal = {Physics of Fluids},     volume = {34},     year = {2022},     url = {https://doi.org/10.1063/5.0097679}, }

Dataset with the node discretisations employed for training advection models in "Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics"

Dataset with the node discretisations employed for training advection models in "Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics" (https://doi.org/10.1063/5.0097679). The training code is available at https://github.com/mario-linov/graphs4cfd.

Bone Age Assessment from Articular Surface and Epiphysis using Deep Neural Networks

This is our research dataset of "Bone Age Assessment from Articular Surface and Epiphysis in Hand Radiography using Deep Neural Networks"

Preprocessed datasets for experiments in the paper "Constrained Monotonic Neural Networks"

Preprocessed datasets used in the experiments of the paper: <ul> <li>Davor Runje, Sharath M. Shankaranarayana. Constrained Monotonic Neural Networks. International Conference on Machine Learning, 2023. [<a href="https://arxiv.org/pdf/2205.11775.pdf">pdf</a>]</li> </ul> The code that runs the experiments can be found at: <a href="https://github.com/airtai/monotonic-dense-layer">https://github.com/airtai/monotonic-dense-layer</a>

Interpretation of EKG with Image Recognition and Convolutional Neural Networks

Dataset used in the training and evaluation of "Interpretation of EKG with Image Recognition and Convolutional Neural Networks" Abstract: Electrocardiograms (EKG) form the backbone of all cardiovascular diagnosis, treatment and follow up. Given the pivotal role it plays in modern medicine, there have been multiple efforts to computerize the EKG interpretation with algorithms to improve efficiency and accuracy. Unfortunately, many of these algorithms are machine specific and run-on proprietary signals generated by that machine, hence not generalizable. We propose the development of an image recognition model which can be used to read standard EKG strips. A convolutional neural network (CNN) was trained to classify 12-lead EKGs between 7 clinically important diagnostic classes. An austere variation of the MobileNetV3 model was trained from the ground up on publicly available labeled training set. The precision per class varies from 52% to 91%. This is a novel approach to EKG interpretation as an image recognition problem.

opencc-by-4.0Mar 2023View details →

Supporting data for "Reliable interpretability of biology-inspired deep neural networks"

Contents data.tgz contains all data necessary for reproducing the analysis in the manuscript. After cloning the GitHub repository, extract the contents of this file into folder data. The archive contains the following subfolders: <ul> <li>dtox DTox results, one subfolder per seed <ul> <li>module_relevance.tsv: contains node importance scores, with the following columns: <ul> <li>(first, unnamed): compound identifier</li> <li>remaining columns: node identifiers (UniProt and Reactome IDs)</li> </ul> </li> <li>test_labels.csv: predictions for the test set, with two columns: <ul> <li>truth: true label (0 or 1)</li> <li>predicted: predicted label (decimal number between 0 and 1)  </li> </ul> </li> </ul> </li> <li>mskimpact_[cancer type]_[experiment] P-NET results using the MSK-IMPACT 2017 dataset, one subfolder per seed [cancer type] is one of bc (breast cancer), cc (colorectal cancer), nsclc (non-small cell lung cancer), or pc (prostate cancer) [experiment] is one of original (original setup) and shuffled (shuffled labels)  </li> <li>pnet_[experiment] P-NET results using the original (prostate cancer) dataset, one subfolder per seed [experiment] is one of deterministic (deterministic input data), original (original setup), and shuffled (shuffled labels) <ul> <li>node_importance.csv: contains node importance scores, with the following columns: <ul> <li>(first, unnamed): node name</li> <li>coef: original node importance scores</li> <li>coef_graph: indegree plus outdegree of node</li> <li>coef_combined: adjusted node importance score (= coef / coef_graph if coef_graph > mean(coef_graph) + 5 sd(coef_graph) in the respective layer)</li> <li>coef_combined_zscore: scaled coef_combined</li> <li>coef_combined2: z(z(coef_graph) - z(coef))</li> <li>layer: layer of the node</li> </ul> </li> <li>predictions_test.csv: predictions for the test set, with the following columns: <ul> <li>(first, unnamed): sample name</li> <li>pred: predicted class (unfortunately, encoded by a double 1.0 or 0.0)</li> <li>pred_scores: probability of the predicted class</li> <li>y: true class (encoded as integer 1 or 0)</li> </ul> </li> <li>predictions_train.csv: predictions for the training set (same columns as above)</li> <li>link_weights_[layer].csv: only in subfolder 234_20080808; matrices with edge weights</li> </ul> </li> </ul>   Changelog v1.1.0  – 2023-06-28 <ul> <li>added DTox results</li> <li>added results of P-NET experiments with MSK-IMPACT 2017 dataset</li> </ul> v1.0.0 – 2023-03-22 <ul> <li>initial release</li> </ul>

opencc-by-4.0Jun 2023View details →

Carbon configurations dataset generated in "A systematic approach to generating accurate neural network potentials: the case of carbon"

This dataset contains 60133 configurations of Carbon as generated in the paper "A systematic approach to generating accurate neural network potentials: the case of carbon". The configurations represent crystal structures containing only Carbon atoms, ranging from 16 to 200 atoms in the unit cell, with energy and forces calculated through density functional theory. Please refer to the original paper for more details about the generation procedure and simulation parameters. The dataset consists of a compressed archive containing a single .xyz file with all configurations, with lattice and energy information contained in the comment line, and one line per atom with positions and forces. Lengths are in Angstroms and energies in eV.

opencc-by-4.0Jul 2023View details →

EuroSAT Model Zoo: A Dataset of Diverse Populations of Neural Network Models - EuroSAT

Abstract 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. Dataset This dataset is part of a larger collection of model zoos and contains the zoos trained on EuroSAT. All zoos with extensive information and code can be found at <a href="http://www.modelzoos.cc">www.modelzoos.cc</a>. This repository contains two types of model populations: the base model zoo ("eurosat_cnn_kaiming_uniform.zip"), as well as a collection of sparsified model zoos (filenames ending in "magn_XX.zip" or "ard.zip"). Zoos are trained with CNN models in configurations varying the seed only (seed), and sparsification is done through magnitude-based weight pruning ("magn_XX.zip") or  varational dropout ("ard.zip"). For more information on the zoos and code to access and use the zoos, please see <a href="http://www.modelzoos.cc">www.modelzoos.cc</a>.

opencc-by-4.0Dec 2022View details →

Multi-Omics Visible Drug Activity Prediction with a Biologically Informed Neural Network Model

Drug discovery is a challenging task, it takes several years for a drug to be introduced on the market, with most of the studied drugs not even passing the first phase. The understanding of the mechanisms influencing response to drugs can reduce failures and accelerate drug development. Virtual drug screening, based on Machine Learning models, is a promising field for the prediction of the outcome of a treatment. However, the complex relationships between the features learned by these models are still poorly understood and not easy to interpret. We have designed a Neural Network model for drug sensitivity prediction that leverages a Visible Neural Network, an easily interpretable model, due to its biologically informed nature. The trained model can be inspected to study which biological processes were fundamental for the prediction and to identify the drug properties that affect sensitivity. It combines multi-omics data from various types of tumor tissues and drug representations based on molecular descriptors. The mechanisms learned from the network can also be exploited to find candidate drugs for synergy to predict the effect of combined therapies. We consider the unbalanced nature of public drug screening datasets and show that our model outperforms state-of-the-art visible machine learning models.

opencc-by-3.0-usJul 2023View details →

CePNEM model analysis data and ANTSUN and microscopy neural network weights

Citation and publication To cite this work or access the publication, please use the citation information listed here: <a href="https://github.com/flavell-lab/AtanasKim-Cell2023/tree/main#citation">https://github.com/flavell-lab/AtanasKim-Cell2023/tree/main#citation</a>   Initially published as preprint in: Brain-wide representations of behavior spanning multiple timescales and states in C. elegans Adam A. Atanas*, Jungsoo Kim*, Ziyu Wang, Eric Bueno, McCoy Becker, Di Kang, Jungyeon Park, Cassi Estrem, Talya S. Kramer, Saba Baskoylu, Vikash K. Mansingkha, Steven W. Flavell bioRxiv 2022.11.11.516186; doi: <a href="https://doi.org/10.1101/2022.11.11.516186">https://doi.org/10.1101/2022.11.11.516186</a> * Equal Contribution   Contents 1. deepnet-weights.tar.bz2 contains the trained weights of the neural networks used in this project. 3dunet_540nm_voxels: 3D U-Net for segmenting neurons head_detector_unet: finding worm head landmark used in ANTSUN registration head_detector_unet_0622: an alternative version of the above, optimal for NeuroPAL datasets microscope_tracker: detecting keypoints for online tracking on the microscope behavior_nir: segmentation of the recorded NIR behavior images for behavior quantification 2. data files ANTSUN processed datasets and CePNEM processed model fits and analysis data. Check the project packages and notebooks in the project github repository (<a href="https://github.com/flavell-lab/AtanasKim-Cell2023/">https://github.com/flavell-lab/AtanasKim-Cell2023/</a>) on using these datasets.

Supplementary dataset for "Evaluating Graph Neural Networks for Link Prediction: Current Pitfalls and New Benchmarking"

The supplementary dataset for the paper "Evaluating Graph Neural Networks for Link Prediction: Current Pitfalls and New Benchmarking". We include the splits for cora, citeseer, and pubmed, the hard negative samples, and the node2vec embeddings. We also include a jupyter file read_data.ipynb to show how to read the non-txt file. <ul> <li>heart_test_samples.npy, heart_valid_samples.npy: the heard negative samples</li> <li>*-n2v-embedding.pt: node2vec embeddings</li> <li>test_samples_index.pt, valid_samples_index.pt: the node index of the selected samples in ogbl-ppa under HeaRT</li> <li>gnn_feature: the input feature of cora, citeseer, pubmed</li> </ul> More details for our code and how to use the dataset are on the code repository: https://github.com/Juanhui28/HeaRT .

Datasets for Paper "BenchTemp: A General Benchmark for Evaluating Temporal Graph Neural Networks"

Datasets for Paper "BenchTemp: A General Benchmark for Evaluating Temporal Graph Neural Networks" URL: https://github.com/qianghuangwhu/benchtemp Openreview: https://openreview.net/forum?id=rnZm2vQq31 There are 19 (15+4)  benchmark temporal graph datasets: reddit, wikipedia, mooc, lastfm, enron, SocialEvo, uci, CollegeMsg, TaobaoSmall, CanParl, Contacts, Flights, UNtrade, USLegis, UNvote, DGraphFin, TaobaoLarge, YoutubeReddit, YoutubeRedditLarge   Each dataset has three files: 1. ml_{data_name}.csv - the csv file of the Temporal Graph. This file have five columns with properties: 'u': the id of the user. 'i': the id of the item. 'ts': the timestamp of the interaction (edge) between the user and the item. 'label': the label of the interaction (edge). 'idx': the index of the interaction (edge). For example: ,u,i,ts,label,idx 0,1,2,0.0,0.0,1 1,1,3,0.0,0.0,2 2,1,4,0.0,0.0,3 2. ml_{data_name}.npy - the edge features corresponding to the interactions (edges) in the the Temporal Graph.. 3. ml_{data_name}_node.npy - the initialization node features of the Temporal Graph.

Prediction of Spheroid Cell Death using Fluorescence Staining and Convolutional Neural Networks

This repository contains training, validation, and testing of fluorescence image data sets with their label for spheroid cell death classification.  These data are intended to be used in the paper "Prediction of Spheroid Cell Death using Fluorescence Staining and Convolutional Neural Networks" currently submitted

Photometric Completeness Modelled With Neural Networks

Neural networks associated with the paper "Photometric Completeness Modelled With Neural Networks" (Harris & Speagle 2023). Neural networks (`nn_clf_[...].joblib`) are included for all possible parameter combinations and trained over various numbers of artificial star tests (`ngc[...].dat`). See the example notebook (`nn_example.ipynb`) for detailed explanations of the files, their contents, and some usage examples.

opencc-by-4.0Oct 2023View details →

Dataset: Parameter estimation by learning quantum correlations in continuous photon-counting data using neural networks

Dataset for the paper E. Rinaldi, M. González Lastre, S. García Herreros, S. Ahmed, M. Khanahmadi, F. Nori, and C. Sánchez Muñoz (2023), <a href="https://arxiv.org/abs/2310.02309">¨Parameter estimation by learning quantum correlations in continuous photon-counting data using neural networks¨, arxiv: 2310.02309</a> This dataset can be used to populate the [datapath] folder in the repository ParamEst-NN (<a href="https://github.com/CarlosSMWolff/ParamEst-NN">github.com/CarlosSMWolff/ParamEst-NN</a> ) and reproduce the results shown in the paper. The dataset consist of four folders: <ol> <li>Training trajectories. Records of quantum-jump trajectories simulated with the Monte-Carlo solver of the <a href="https://qutip.org/">QuTiP</a> library, used to train neural networks for the problem of quantum parameter estimation. The records consist of time delays between quantum jumps.</li> <li>Models. Models trained with the training trajectories provided, and used to obtain the results shown in the paper.</li> <li>Validation trajectories. Trajectories used to benchmark the trained models. For the 2D case, the same trajectories are provided as a single .npy file, and split in 10 separated batches inside a ```batches``` folder. These are the batches that we used to generate Bayesian estimations on a cluster using nested sampling (see README file of the <a href="http://github.com/CarlosSMWolff/ParamEst-NN">repository</a>).</li> <li>Cached results. Here we provide pre-computed Bayesian estimations for the 2D multi-parameter estimation case using nested sampling.</li> </ol>   E.R. was supported by Nippon Telegraph and Telephone Corporation (NTT) Research during the early stages of this work. C.S.M. acknowledges that the project that gave rise to these results received the support of a fellowship from “la Caixa” Foundation (ID 100010434) and from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie Grant Agreement No.847648, with fellowship code LCF/BQ/PI20/11760026, and financial support from the MCINN project PID2021-126964OB-I00 (QENIGMA) and the Proyecto Sinérgico CAM 2020 Y2020/TCS- 6545 (NanoQuCo-CM).

UKESM1.0-ice simulation output used as test data in Burgard et al., Emulating present and future simulations of melt rates at the base of Antarctic ice shelves with neural networks

These files contain NEMO ocean model output and domain definitions for the Southern Ocean from UKESM1.0-ice simulations described in section 6.3.2 of Smith et al. "Coupling the U.K. Earth System Model to Dynamic Models of the Greenland and Antarctic Ice Sheets" , Journal of Advances in Modeling Earth Systems, 2021 Files labelled "bf663" are the UKESM simulation referred to in that section as "constant 1970 greenhouse gas and other forcings". Files labelled bi646 are the UKESM simulation referred to in that section as "instantaneously quadrupled 1970 CO2 concentrations". They were used as test data for the performance of neural networks in Burgard et al., "Emulating present and future simulations of melt rates at the base of Antarctic ice shelves with neural networks", Journal of Advances in Modeling Earth Systems 2023.