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.
69
datasets available to search
ShareScore release 0.9.0
Dataset results
69 results for “Representation learning”
Unlabeled Sentinel 2 time series dataset (training, T31UEP): Self-Supervised Spatio-Temporal Representation Learning of Satellite Image Time Series
<p>This is a part of the unlabeled Sentinel 2 (S2) L2A dataset composed of patch time series acquired over France used to pretrain U-BARN. For further details, see section IV.A of the pre-print article "Self-Supervised Spatio-Temporal Representation Learning Of Satellite Image Time Series" available <a href="https://hal.science/hal-04084839">here</a>. Each patch is constituted of the 10 bands [B2,B3,B4,B5,B6,B7,B8,B8A,B11,B12] and the three masks ['CLM_R1', 'EDG_R1', 'SAT_R1']. The global dataset is composed of two disjoint datasets: training (9 tiles) and validation dataset (4 tiles).</p> <p>In this repo,<strong> only data from the S2 tile T31UEP</strong> are available. To download the full pretraining dataset, see: <a href="https://doi.org/10.5281/zenodo.7891924">10.5281/zenodo.7891924</a></p> <table> <caption><strong>Global unlabeled dataset description</strong></caption> <tbody> <tr> <td>Dataset name</td> <td>S2 tiles</td> <td>ROI size</td> <td>Temporal extent</td> </tr> <tr> <td>Train</td> <td> <p>T30TXT,T30TYQ,T30TYS,T30UVU,</p> <p>T31TDJ,T31TDL,T31TFN,T31TGJ,T31UEP</p> </td> <td>1024*1024</td> <td>2018-2020</td> </tr> <tr> <td>Val</td> <td>T30TYR,T30UWU,T31TEK,T31UER</td> <td>256*256</td> <td>2016-2019</td> </tr> </tbody> </table>
BISCUIT: Causal Representation Learning from Binary Interactions
<p>This repository contains the datasets from the paper "BISCUIT: Causal Representation Learning from Binary Interactions" (<a href="https://phlippe.github.io/BISCUIT/">link</a>) by Phillip Lippe, Sara Magliacane, Sindy Löwe, Yuki M. Asano, Taco Cohen, Efstratios Gavves. </p> <p><strong>iTHOR Embodied AI </strong>- The Embodied AI dataset is generated with the iTHOR simulator. We use the default kitchen environment, FloorPlan10, and position the robot in front of the kitchen counter. The robot interacts with different objects in the environment, including a Microwave, cabinet, stove, and an egg. For more details on the dataset, see <a href="https://github.com/phlippe/BISCUIT">our GitHub repository</a> and the appendix of our paper.</p> <p><strong>CausalWorld </strong>- The CausalWorld environment implements a tri-finger robot which can interact with a cube in the center of a stage. We additionally introduce causal variables for the friction parameters of the stage, floor and cube, as well as color changes. For more details on the dataset as well as the code to generate this dataset, see <a href="https://github.com/phlippe/BISCUIT">our GitHub repository</a> and the appendix of our paper.</p> <p><strong>Voronoi</strong> - The Voronoi benchmark allows for creating causal systems with arbitrary number of causal variables and causal graphs. We provide datasets with 6 and 9 variables, as well as systems with minimal number of interactions. For details, see <a href="https://github.com/phlippe/BISCUIT">our GitHub repository</a> and our paper.</p>
The neural correlates of embodied L2 learning Does embodied L2 verb learning affect representation and retention?
<p>We investigated how naturalistic actions in a highly immersive, multimodal, interactive 3D virtual reality (VR) environment may enhance word encoding by recording EEG in a pre/post-test learning paradigm. While behavior data has shown that coupling word encoding with gestures congruent with word meaning enhances learning, the neural underpinnings of this effect have yet to be elucidated. We coupled EEG recording with VR to examine whether “embodied learning” improves learning and creates linguistic representations that produce greater motor resonance. Participants learned action verbs in an L2 in two different conditions: Specific action (observing and performing congruent actions on virtual objects) and Pointing (observing actions and pointing to virtual objects). Pre and post-training participants performed a Match-mismatch task as we measured EEG (variation in the N400 response as a function of match between observed actions and auditory verbs) and a Passive listening task while we measured motor activation (mu (8-13 Hz) and beta band (13-30Hz) desynchronization during auditory verb processing) during verb processing. Contrary to our expectations, post-training results revealed neither semantic nor motor effects in either group when considered independently of learning success. Behavioral results showed both groups learned the verbs, but also a great deal of variability in learning success. When considering performance, Low performance learners showed no semantic effect and High performance learners exhibited an N400 effect for Mismatch vs Match trails post-training, independent of the type of learning. Taken as a whole, our results suggest that embodied processes can play an important role in L2 learning.</p>
Data from: Learning enhances behaviorally relevant representations in apical dendrites
Open the record for dataset details and reuse information.
Learning Unsupervised Knowledge-Enhanced Representations to Reduce the Semantic Gap in Information Retrieval (Evaluation datasets)
<p>This dataset contains all the runs, pools, plots and analyses to reproduce the results presented in the paper: "Learning Unsupervised Knowledge-Enhanced Representations to Reduce the Semantic Gap in Information Retrieval ", 2020.</p>
Experimental Data for the Paper 'Rotation-Aware Representation Learning for Remote Sensing Image Retrieval'
<p><strong>Experimental Data for the Paper 'Rotation-Aware Representation Learning for Remote Sensing Image Retrieval'</strong></p> <p>In this repository, we provide the implementation of the algorithms developed in the paper 'Rotation-Aware Representation Learning for Remote Sensing Image Retrieval' along with the experimental results.<br> The goal is to provide the elements needed to validate and reproduce our research work as well as all the tools needed to reach the same conclusions as we did.<br> The licences valid for the elements of this repository are discussed under point "2. Licenses" below.</p> <p><em><strong>1. Structure</strong></em></p> <p>The repository contains the following items:</p> <ol> <li>"data" - the results from our experiments</li> <li>"lib" - some external functions used in the experiments</li> <li>"make_data" - the training and test data</li> <li>"fmt-vgg.py" - the FMT-RAN model</li> <li>"stn.py" - the STN module of ST-RAN</li> <li>"st_ran.py" - the ST-RAN model</li> <li>"README" - this text here.</li> <li>"LICENSE" - the <a href="https://mit-license.org/">MIT License</a></li> </ol> <p><strong><em>2. License</em></strong></p> <p>The following licenses apply for the files and folders:</p> <ul> <li>The files "stn.py" and "spatial_transformer_tutorial.py" in the folder "lib" are from the GitHub repository <a href="https://github.com/GHamrouni/stn-tuto">https://github.com/GHamrouni/stn-tuto</a> and therefore are under the copyright of its repository owner Ghassen Hamrouni.</li> <li>All other files are under the <a href="https://mit-license.org/">MIT License</a>.</li> </ul> <p>The <a href="https://mit-license.org/">MIT License</a> is included here as file "LICENSE".</p> <p><em><strong>3. Contact</strong></em></p> <p>1. Dr. <a href="http://iao.hfuu.edu.cn/146">Zhize WU</a>, <a href="mailto:wuzz@hfuu.edu.cn">wuzz@hfuu.edu.cn</a><br> 2. Dr. <a href="http://iao.hfuu.edu.cn/5">Thomas WEISE</a>, <a href="mailto:tweise@hfuu.edu.cn">tweise@hfuu.edu.cn</a>, <a href="http://mailto:tweise@ustc.edu.cn">tweise@ustc.edu.cn</a></p> <p><a href="http://iao.hfuu.edu.cn">Institute of Applied Optimization</a>, <br> School of Artificial Intelligence and Big Data, <br> Hefei University, South Campus 2, Jinxiu Dadao 99, <br> Hefei Economic and Technological Development Area, <br> Shushan District, Hefei 230601, Anhui, China</p>
The neural correlates of embodied L2 learning. Does embodied L2 verb learning affect representation and retention?
<p>We investigated how naturalistic actions in a highly immersive, multimodal, interactive 3D virtual reality (VR) environment may enhance word encoding by recording EEG in a pre/post-test learning paradigm. While behavior data has shown that coupling word encoding with gestures congruent with word meaning enhances learning, the neural underpinnings of this effect have yet to be elucidated. We coupled EEG recording with VR to examine whether "embodied learning" improves learning and creates linguistic representations that produce greater motor resonance. Participants learned action verbs in an L2 in two different conditions: Specific action (observing and performing congruent actions on virtual objects) and Pointing (observing actions and pointing to virtual objects). Pre and post-training participants performed a Match-mismatch task as we measured EEG (variation in the N400 response as a function of match between observed actions and auditory verbs) and a Passive listening task while we measured motor activation (mu (8-13 Hz) and beta band (13-30Hz) desynchronization during auditory verb processing) during verb processing. Contrary to our expectations, post-training results revealed neither semantic nor motor effects in either group when considered independently of learning success. Behavioral results showed both groups learned the verbs, but also a great deal of variability in learning success. When considering performance, Low performance learners showed no semantic effect and High performance learners exhibited an N400 effect for Mismatch vs Match trails post-training, independent of the type of learning. Taken as a whole, our results suggest that embodied processes can play an important role in L2 learning.</p>
Reactivation of hedonic but not sensory representations in human emotional learning
<p>This repository contains data used in the following papers</p> <p>Reactivation of hedonic but not sensory representations in human emotional learning<br> Ehlers MR, Kryklywy JH, Beukers AO, Moore SR, Forys BJ, Anderson AK, Todd RM<br> https://doi.org/10.1101/2021.11.25.469891</p>
Data for COLLAPSE: A representation learning framework for identification and characterization of protein structural sites
<p>Data for methods described in the paper "COLLAPSE: A representation learning framework for identification and characterization of protein structural sites" by Alexander Derry and Russ B. Altman (BioRxiv, 2022). https://www.biorxiv.org/content/10.1101/2022.07.20.500713v1</p>
Representation learning for multi-modal spatially resolved transcriptomics data
<p>This folder contains the already unified input used for the models. The raw data is referenced here:</p> <ul> <li>LIBD Human DLPFC dataset is available at <a href="https://github.com/LieberInstitute/HumanPilot">https://github.com/LieberInstitute/HumanPilot</a> and <a href="http://research.libd.org/spatialLIBD" rel="nofollow">http://research.libd.org/spatialLIBD</a>;</li> <li>Human Breast Cancer - Zenodo <a href="https://doi.org/10.5281/zenodo.4739739" rel="nofollow">https://doi.org/10.5281/zenodo.4739739</a>,</li> <li>Human Liver Normal and Cancer - <a href="https://nanostring.com/products/cosmx-spatial-molecular-imager/human-liver-rna-ffpe-dataset/" rel="nofollow">https://nanostring.com/products/cosmx-spatial-molecular-imager/human-liver-rna-ffpe-dataset/</a>.</li> </ul>
Federated Learning Games for Reconfigurable Intelligent Surfaces via Causal Representations: SW and Data
<p>This upload contains the main simulation code and related datasets used in the conference paper entitled <a href="https://ieeexplore.ieee.org/document/10437657" target="_blank" rel="nofollow noreferrer noopener">Federated Learning Games for Reconfigurable Intelligent Surfaces via Causal Representations</a>, which was presented at <a href="https://globecom2023.ieee-globecom.org/" target="_blank" rel="nofollow noreferrer noopener">IEEE GLOBECOM 2023</a>.</p>
Data and Code: Cortical representations of symbolic and non-symbolic quantity expand but become estranged with learning and development
<h1><strong>Note</strong></h1> <p>Here we provide preprocessed data and analysis code used in "Cortical representations of symbolic and non-symbolic quantity expand but become estranged with learning and development".</p> <p>Because of anonymization concerns within the framework of EU privacy regulations (<a href="https://gdpr-info.eu/">GDPR</a>), we cannot provide raw MRI data. Therefore, the fMRI data consists of individual preprocessed volumes, normalized into the MNI template, and averaged across five TRs for each block (see paper for details about the preprocessing pipeline).</p> <p>The analysis code requires Python version 3.8.8, Nilearn version 0.8.1, and Scikit-learn version 0.24.1.</p> <p>If you have any questions, please send an email to nakai.tomoya [at] neuro.mimoza.jp. </p> <p> </p> <h1><strong>Usage</strong></h1> <pre>import PredysDecoding5ans_SearchLight_LOOCV as pdsl5 import PredysDecoding8ans_SearchLight_LOOCV as pdsl8 import PredysDecoding5to8_SearchLight as pdsl58 import PredysDecoding8to5_SearchLight as pdsl85 </pre> <h3>Within-format decoding for 5-year-olds (Figures 2A, B):</h3> <pre>pdsl5.IntraModalDec(TaskName='Dots') pdsl5.SaveNifti_PermTest(TaskName='Dots') pdsl5.IntraModalDec(TaskName='Digits') pdsl5.SaveNifti_PermTest(TaskName='Digits')</pre> <h3>Within-format decoding for 8-year-olds (Figure 2C, D):</h3> <pre>pdsl8.IntraModalDec(TaskName='Dots') pdsl8.SaveNifti_PermTest(TaskName='Dots') pdsl8.IntraModalDec(TaskName='Digits') pdsl8.SaveNifti_PermTest(TaskName='Digits')</pre> <h3>Within-format decoding, paired tests between 5- and 8-year-olds (Figures 3A, B):</h3> <pre>pdsl5.SaveNifti_Paired_PermTest(TaskName='Dots') pdsl5.SaveNifti_Paired_PermTest(TaskName='Digits') pdsl8.SaveNifti_Paired_PermTest(TaskName='Dots') pdsl8.SaveNifti_Paired_PermTest(TaskName='Digits')</pre> <h3>Within-format decoding across 5- and 8-year-olds (Figures 3C, D):</h3> <pre>pdsl85.IntraModalDec(TaskName='Dots') pdsl85.SaveNifti_PermTest(TaskName='Dots') pdsl85.IntraModalDec(TaskName='Digits') pdsl85.SaveNifti_PermTest(TaskName='Digits') pdsl58.IntraModalDec(TaskName='Dots') pdsl58.SaveNifti_PermTest(TaskName='Dots') pdsl58.IntraModalDec(TaskName='Digits') pdsl58.SaveNifti_PermTest(TaskName='Digits') pdsl58.SaveNifti_PermTest_Conj(TaskName='Dots') pdsl58.SaveNifti_PermTest_Conj(TaskName='Digits')</pre> <h3>Between-format decoding for 5-year-olds (Figures 4A, 5B):</h3> <pre>pdsl5.CrossModalDec(TaskName1='Dots', TaskName2='Digits') pdsl5.SaveNifti_PermTest(TaskName='Dots2Digits') pdsl5.CrossModalDec(TaskName1='Digits', TaskName2='Dots') pdsl5.SaveNifti_PermTest(TaskName='Digits2Dots') pdsl5.SaveNifti_PermTest_Conj(TaskName1='Dots2Digits', TaskName2='Digits2Dots') pdsl5.CrossModalDec(TaskName1='Dots', TaskName2='Letters') pdsl5.SaveNifti_PermTest(TaskName='Dots2Letters') pdsl5.CrossModalDec(TaskName1='Letters', TaskName2='Dots') pdsl5.SaveNifti_PermTest(TaskName='Letters2Dots') pdsl5.SaveNifti_PermTest_Conj(TaskName1='Dots2Letters', TaskName2='Letters2Dots')</pre> <h3>Between-format decoding for 8-year-olds (Figure 4A):</h3> <pre>pdsl8.CrossModalDec(TaskName1='Dots', TaskName2='Digits') pdsl8.SaveNifti_PermTest(TaskName='Dots2Digits') pdsl8.CrossModalDec(TaskName1='Digits', TaskName2='Dots') pdsl8.SaveNifti_PermTest(TaskName='Digits2Dots') pdsl8.SaveNifti_PermTest_Conj(TaskName1='Dots2Digits', TaskName2='Digits2Dots')</pre> <h3>Between-format decoding, paired tests between 5- and 8-year-olds (Figure 4B)</h3> <pre>pdsl5.SaveNifti_Paired_PermTest_Conj(TaskName1='Dots2Digits', TaskName2='Digits2Dots') pdsl8.SaveNifti_Paired_PermTest_Conj(TaskName1='Dots2Digits', TaskName2='Digits2Dots')</pre> <p> </p>
Distinct representations of innate and learned threats within the thalamic-amygdala pathway
<p>Behavioral flexibility and timely reactions to salient stimuli are essential for survival. The subcortical thalamic-basolateral amygdala (BLA) pathway serves as a shortcut for salient stimuli ensuring rapid processing. Here, we show that BLA neuronal and thalamic axonal activity mirror the defensive behavior evoked by an innate visual threat as well as an auditory learned threat. Importantly, perturbing this pathway compromises defensive responses to both forms of threats, in that animals fail to switch from exploratory to defensive behavior. Despite the shared pathway between the two forms of threat processing, we observed noticeable differences. Blocking beta-adrenergic receptors impair the defensive response to the innate but not the learned threats. This reduced defensive response, surprisingly, is reflected in the suppression of the activity exclusively in the BLA, as the thalamic input response remains intact. Our side-by-side examination highlights the similarities and differences between innate and learned threat-processing, thus providing new fundamental insights.</p>
Multi-Scale Representation Learning on Proteins
<p>Protein-ligand structures from PDBBind (v2019), and protein structures from the enzyme dataset as described in the paper: "Multi-Scale Representation Learning on Proteins" with associated code at https://github.com/vsomnath/holoprot.</p> <p><strong>Paper Abstract</strong></p> <p>Proteins are fundamental biological entities mediating key roles in cellular function and disease. This paper introduces a multi-scale graph construction of a protein – HoloProt – connecting surface to structure and sequence. The surface captures coarser details of the protein, while sequence as primary component and structure – comprising secondary and tertiary components – capture finer details. Our graph encoder then learns a multi-scale representation by allowing each level to integrate the encoding from level(s) below with the graph at that level. We test the learned representation on different tasks, (i.) ligand binding affinity (regression), and (ii.) protein function prediction (classification). On the regression task, contrary to previous methods, our model performs consistently and reliably across different dataset splits, outperforming all baselines on most splits. On the classification task, it achieves a performance close to the top-performing model while using 10x fewer parameters. To improve the memory efficiency of our construction, we segment the multiplex protein surface manifold into molecular superpixels and substitute the surface with these superpixels at little to no performance loss.</p>
Distinct representations of innate and learned threats within the thalamic-amygdala pathway
Open the record for dataset details and reuse information.
PhyloCNN: Improving tree representation and neural network architecture for deep learning from trees in phylodynamics and diversification studies
Open the record for dataset details and reuse information.
Learning Context-aware Structural Representations to Predict Antigen and Antibody Binding Interfaces
<p>These are data corresponding to the paper '<a href="https://doi.org/10.1093/bioinformatics/btaa263">Learning Context-aware Structural Representations to Predict Antigen and Antibody Binding Interfaces</a>', published in Bioinformatics 2020. The code to run the experiments can be found on github <a href="https://github.com/vamships/PECAN">here</a>.</p>
[Data] Self-Supervised Bayesian Representation Learning of Acoustic Emissions from Laser Powder Bed Fusion Process for In-situ Monitoring
<div> <div> <div> <p>Different Laser Powder Bed Fusion (LPBF) process spaces were deliberately introduced by employing two distinct 316L stainless steel powder distributions (with particle sizes >45 μm and < 45 μm) and processing them with two sets of laser parameters, resulting in the creation of four datasets [D1, D2, D3, and D4]. These datasets encompass LoF pores, conduction mode, and keyhole formations, each associated with three LPBF regimes denoted as D1, D2, D3, and D4. The experiments utilized a Sisma MYSINT 100 commercial LPBF printer and an airborne AE sensor system with a flat frequency response ranging from 0 to 150 kHz. Validation of the ground truths for the three laser regimes across the four datasets, representing distinct process spaces, was accomplished through the confirmation of cross-sectional images. In the course of fabricating a cube using a powder bed and laser, data acquisition from an AE sensor was triggered when the optical intensity reached a threshold of 0.5 V for each scan length. The photodiode trigger gain was adjusted to saturate at 5 V, and the ensuing continuous-time window, where the optical signal remained at 5 V for 12.5 ms, was calculated and segmented to generate the dataset. Irrespective of the specific regime (Lack of Fusion, Conduction, and Keyhole) or the cube being fabricated (with two powder distributions), the signals obtained during this process were then segmented into a 12.5 ms window comprising 5000 data points. To eliminate any noise, an offline application of a low-pass Butterworth filter with a 150 kHz cut-off frequency was employed, aligned with the frequency response specification of the AE sensor. Each dataset has two files against it [raw/groundtruth label].</p> </div> </div> </div>
Learning Useful Representations of Recurrent Neural Network Weight Matrices
<p>Dataset of RNN weights for the ICML 2024 paper "Learning Useful Representations of Recurrent Neural Network Weight Matrices". See <a title="GitHub repository" href="https://github.com/vincentherrmann/rnn-weights-representation-learning">https://github.com/vincentherrmann/rnn-weights-representation-learning</a>.</p> <h1> </h1>
Characterization of tumour heterogeneity through segmentation-free representation learning on multiplexed imaging data
<p>This is the data repository for Characterization of tumour heterogeneity through segmentation-free representation learning on multiplexed imaging data.</p> <p>Catalog:</p> <ol> <li>Intermediate data used in plotting: CANVAS_source_data.zip <ol> <li>Single cell monocyte data: monocyte.h5ad</li> <li>qPCR table: qPCR_1013.csv</li> <li>NanoString GeoMx cell composition: fig6b.csv</li> </ol> </li> <li>Pretrained CANVAS model: checkpoint-1999.pth</li> </ol> <p>The source IMC data is avaiable at: https://zenodo.org/records/7760826</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.