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.
1,952
datasets available to search
ShareScore release 0.9.0
Dataset results
1,952 results for “Action”
The Machine-Actionable Ancient Text (MAAT) Corpus
<p>The Machine-Actionable Ancient Text (MAAT) Corpus is a new resource providing training and evaluation data for restoring lacunae in ancient Greek, Latin, and Coptic texts. Current text restoration systems require large amounts of data for training and task-relevant means for evaluation. The MAAT Corpus addresses this need by converting texts available in EpiDoc XML format into a machine-actionable format that preserves the most textually salient aspects needed for machine learning: the text itself, unclear letters, restorations, and lacunae. Structured test cases are generated from the corpus that align with the actual text restoration task performed by papyrologists and epigraphist, enabling more realistic evaluation than the synthetic tasks used previously. The initial 1.0 beta release contains approximately 134,000 text editions, 178,000 text blocks, and 750,000 individual restorations, with Greek and Latin predominating. This corpus aims to facilitate the development of computational methods to assist scholars in accurately restoring ancient texts.</p>
Quantitative account of social interactions in a mental health care ecosystem: cooperation, trust and collective action
<p>Mental disorders have an enormous impact in our society, both in personal terms and in the economic costs associated with their treatment. In order to scale up services and bring down costs, administrations are starting to promote social interactions as key to care provision. We analyze quantitatively the importance of communities for effective mental health care, considering all community members involved. By means of citizen science practices, we have designed a suite of games that allow to probe into different behavioral traits of the role groups of the ecosystem. The evidence reinforces the idea of community social capital, with caregivers and professionals playing a leading role. Yet, the cost of collective action is mainly supported by individuals with a mental condition - which unveils their vulnerability. The results are in general agreement with previous findings but, since we broaden the perspective of previous studies, we are also able to find marked differences in the social behavior of certain groups of mental disorders. We finally point to the conditions under which cooperation among members of the ecosystem is better sustained, suggesting how virtuous cycles of inclusion and participation can be promoted in a ’care in the community’ framework.</p>
Blue-Action IPSL AMOC 26N reconstruction
<p>Reconstruction of the past variability of the AMOC at 26°N using the IPSL-CM6A-LR climate model and a surface nudging of ERSST and Friedman et al (2017) SSS.</p>
The DARRL dataset: Demonstrations for Action Recognition and Robot Learning
<p>The DARRL dataset (Demonstrations for Action Recognition and Robot Learning) is a collection of 760 RGB-D videos of humans performing various manipulation tasks. It is provided with object and action annotations (in the COCO format) for 30 of those videos; segmentation masks are also provided.</p> <p>It can also be used as a basis for learning from demonstrations for a robotic arm, for instance.</p> <p> </p> <p>This work is supported by Région Pays de la Loire.</p>
Data set accompanying the research article "Complete representation of action space and value in all striatal pathways"
<p>GCaMP6s calcium imaging data set recorded from freely behaving mice performing open field and 2-choice decision-making tasks using miniscopes. Mice were implanted in the right dorsomedial striatum and three types of output neurons were genetically targeted using transgenic Cre-lines. The data set comprises single-cell spatial filters and calcium activity traces extracted using CaImAn (https://github.com/flatironinstitute/CaImAn) as well as behavioral event logs and tracking coordinates. For more details please refer to the article "Complete representation of action space and value in all striatal pathways" published by the data sets' authors. Analysis code can be found at https://doi.org/10.5281/zenodo.5034618.</p>
Outsourcing asylum reception Street-level organizations and the privatization of state action in Switzerland
<p>By approaching outsourcing through an ethnographic lens,we focus on the policies that enable public authorities to deal with the many uncertainties of refugee reception. We investigate how cantonal (subnational) governments in Switzerland remodel these uncertainties by implementing the policy of reception through private intermediaries, namely actors, instruments, and rationalities from the private sphere.</p> <p>Codebook along DDI standard in pdf and xml formats.</p>
Data and code for "Carbon neutrality should not be the end goal: Lessons for institutional climate action from U.S. higher education"
<p>Code and data for the paper "Carbon neutrality should not be the end goal: Lessons for institutional climate action from U.S. higher education"</p> <p>File descriptions:</p> <p>'HEI_analysis_OneEarth.Rmd' is the code with improved annotation and colorblind-friendly figures.</p> <p>All other data files are provided as excel and csv for convenience.</p> <p>'working_master_data' contains data from the Second Nature reporting platform on emissions by category for each institution analyzed in the paper (measured in metric tons). All adjustments necessary to fill in the data gaps in this file are documented at the beginning of 'HEI_analysis'.</p> <p>'offsets' contains data on the type(s) of offsets purchased by each school in their carbon neutral year (measured in metric tons). This data was assembled from a variety of sources which are documented at the beginning of 'HEI_analysis'.</p> <p>'carbon_neutral_years' contains yearly counts of higher education neutrality goals that were reported to Second Nature as of November 2020.</p>
Water analysis dataset > Water Sentinels pilot project - ACTION project
<p><a href="http://www.ocean-alive.org/en/water-sentinels">WATER SENTINELS</a> is a citizen-science pilot by Ocean Alive to promote water quality at Sado estuary.</p> <p>This dataset presents data and metadata for water quality assessment on water samples collected at Sado estuary by citizens during the 6-month pilot. The samples were analyzed by students and researchers at Setúbal School of Technology (Polytechnic Institute of Setúbal). </p> <p>The data collected by citizens at the field was location (GPS coordinates), temperature and water transparency (using the secchi disc method). In the laboratory researchers evaluated pH, salinity, conductivity, and concentration of ammonia, nitrate, phosphate, suspended solids (SST) and organic matter (SOM and DOM).</p> <p>WATER SENTINELS is one of the pilots supported by <a href="https://actionproject.eu/water-sentinels/"><strong>ACTION</strong></a> project. This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement number 824603.</p>
A dataset of GitHub Actions workflow histories
<p>This replication package accompagnies the dataset and exploratory empirical analysis reported in the paper "A dataset of GitHub Actions workflow histories" published in the IEEE MSR 2024 conference. (The Jupyter notebook can be found in previous version of this dataset).</p> <p><em><strong>Important notice :</strong> It looks like Zenodo is compressing gzipped files two times without notice, they are "double compressed". So, when you download them they should be named : <code>x.gz.gz</code> instead of <code>x.gz</code>. Notice that the provided MD5 refers to the original file. </em></p> <p><em><strong>2025-10-09 update: update repositories list and observation period</strong>. We now have 3M+ workflows from 49.2K+ repositories. We consider repositories with at least one commit after August 25th, 2024, and they were pulled on August 25th-26th, 2025.</em></p> <p>2025-04-15 update: fix missing metadata and minor notation bug. (unchanged observation period)</p> <p><strong>2024-10-25 update: update repositories list and observation period</strong>. <em>We now have 2.3M+ workflows from 43.3K+ repositories. We consider repositories with at least one commit after January 1st, 2024, and they were pulled on October 7th, 2024.</em></p> <p>2024-07-09 update: fix sometimes invalid <code>valid_yaml</code> flag.</p> <p>2024-04-30: initial version</p> <p>The dataset was created as follow : </p> <ol> <li>First, we used GitHub SEART (on August 25th, 2025) to get a list of every non-fork repositories created at least one year before. having at least 300 commits and at least 100 stars where at least one commit was made in the last year. (The goal of these filter is to exclude experimental and personnal repositories).</li> <li>We checked if a folder <code>.github/workflows</code> existed. We filtered out those that did not contained this folder and pulled the others (on August 25th-26th, 2025).</li> <li>We applied the tool <code>gigawork</code> (version 1.4.2) to extract every files from this folder. The exact command used is <code>python batch.py -d /ourDataFolder/repositories -e /ourDataFolder/errors -o /ourDataFolder/output -r /ourDataFolder/repositories_everything.csv.gz -- -w /ourDataFolder/workflows_auxiliaries</code>. (The script <code>batch.py</code> can be found <a href="https://github.com/cardoeng/gigawork/blob/master/scripts/batch.py" target="_blank" rel="noopener">on GitHub</a>).</li> <li>We concatenated every files in <code>/ourDataFolder/output</code> into a csv (using <code>cat headers.csv output/*.csv > workflows_auxiliaries.csv</code> in <code>/ourDataFolder</code>) and compressed it.</li> <li>We added the column <code>uid</code> via a script available <a href="https://github.com/cardoeng/gigawork/blob/master/scripts/uid.py">on GitHub.</a></li> <li>Finally, we archived the folder with pigz <code>/ourDataFolder/workflows</code> (<code>tar -c --use-compress-program=pigz -f workflows_auxiliaries.tar.gz /ourDataFolder/workflows</code>)</li> </ol> <p>Using the extracted data, the following files were created :</p> <ol> <li><code>workflows.tar.gz</code> contains the dataset of GitHub Actions workflow file histories.</li> <li><code>workflows_auxiliaries.tar.gz</code> is a similar file containing also auxiliary files.</li> <li><code>workflows.csv.gz</code> contains the metadata for the extracted workflow files.</li> <li><code>workflows_auxiliaries.csv.gz</code> is a similar file containing also metadata for auxiliary files.</li> <li><code>repositories.csv.gz</code> contains metadata about the GitHub repositories containing the workflow files. These metadata were extracted using the SEART Search tool. </li> </ol> <p>The metadata is separated in different columns:</p> <ol> <li><code>repository</code>: The repository (author and repository name) from which the workflow was extracted. The separator "/" allows to distinguish between the author and the repository name</li> <li><code>commit_hash</code>: The commit hash returned by git</li> <li><code>author_name</code>: The name of the author that changed this file</li> <li><code>author_email</code>: The email of the author that changed this file</li> <li><code>committer_name</code>: The name of the committer</li> <li><code>committer_email</code>: The email of the committer</li> <li><code>committed_date</code>: The committed date of the commit</li> <li><code>authored_date</code>: The authored date of the commit</li> <li><code>file_path</code>: The path to this file in the repository</li> <li><code>previous_file_path</code>: The path to this file before it has been touched</li> <li><code>file_hash</code>: The name of the related workflow file in the dataset</li> <li><code>previous_file_hash</code>: The name of the related workflow file in the dataset, before it has been touched</li> <li><code>git_change_type</code>: A single letter (A,D, M or R) representing the type of change made to the workflow (Added, Deleted, Modified or Renamed). This letter is given by <code>gitpython</code> and provided as is. </li> <li><code>valid_yaml</code>: A boolean indicating if the file is a valid YAML file.</li> <li><code>probably_workflow</code>: A boolean representing if the file contains the YAML key <code>on</code> and <code>jobs</code>. (Note that it can still be an invalid YAML file).</li> <li><code>valid_workflow</code>: A boolean indicating if the file respect the syntax of GitHub Actions workflow. A freely available JSON Schema (used by gigawork) was used in this goal.</li> <li><code>uid</code>: Unique identifier for a given file surviving modifications and renames. It is generated on the addition of the file and stays the same until the file is deleted. Renamings does not change the identifier.</li> </ol> <p>Both <code>workflows.csv.gz</code> and <code>workflows_auxiliaries.csv.gz</code> are following this format.</p>
On Gorenstein Fano threefolds with an action of a twodimensional torus
<p>A database of the non-toric, <span class="math-tex">\(\mathbb{Q}\)</span>-factorial, log terminal, Gorenstein, Fano threefolds of Picard number 1 that admit an effective action of a two-dimensional torus.</p> <p>For more information on the provided data see <a href="https://arxiv.org/abs/2108.03029">arXiv:2108.03029</a>. This dataset is also available at <a href="https://github.com/abaeuerle/fano-3d-lt-gor-rho-1">Github</a>.</p> <p><strong>Description:</strong></p> <p>The columns of the csv-file are labelled as follows:</p> <ul> <li><strong>ID:</strong> The ID of the family within the database, numbered 1 through 538.</li> <li><strong>case:</strong> Refers to the corresponding case in Proposition 2.24. Numbered 1 through 8.</li> <li><strong>format:</strong> The format of the family as defined on page 13.</li> <li><strong>generator matrix:</strong> The defining matrix <span class="math-tex">\(P\)</span> as defined in Construction 2.2.</li> <li><strong>class group: </strong>The isomorphy type of the class group. Presented as <span class="math-tex">\([0, t_1, \dots, t_n]\)</span> meaning <span class="math-tex">\(\mathbb{Z} \oplus ( \mathbb{Z}/t_1 \mathbb{Z} \oplus \dots \oplus \mathbb{Z}/t_n \mathbb{Z} )\)</span>.</li> <li><strong>gd matrix:</strong> Grading matrix of the Cox ring. The <span class="math-tex">\(i\)</span>-th column is the degree of the <span class="math-tex">\(i\)</span>-th Cox ring generator.</li> <li><strong>relations:</strong> Minimal generating set of the ideal of relations of the Cox ring.</li> <li><strong>relation degree:</strong> Degree of the minimal generators of the ideal of relations.</li> <li><strong>anticanonical class:</strong> The anticanonical class.</li> <li><strong>degree:</strong> Anticanonical self-intersection number <span class="math-tex">\(-\mathcal{K}_X^3\)</span>.</li> <li><strong>gorenstein index:</strong> The smallest positive <span class="math-tex">\(\iota \in \mathbb{Z}\)</span> such that <span class="math-tex">\(\iota \mathcal{K}_X\)</span> is Cartier. Here always equal to 1.</li> <li><strong>fano index:</strong> The largest <span class="math-tex">\(k \in \mathbb{Z}\)</span> such that <span class="math-tex">\(k D = \mathcal{K}_X\)</span> for some Cartier divisor <span class="math-tex">\(D\)</span>.</li> <li><strong>dimension:</strong> The dimension of the corresponding variety. Here always equal to 3.</li> <li><strong>picard rank:</strong> The rank of the Picard group. Here always equal to 1.</li> <li><strong>acr gd matrix:</strong> Grading matrix of the anticanonical ring w.r.t. to a minimal system of generators. The <span class="math-tex">\(i\)</span>-th column is the degree of the <span class="math-tex">\(i\)</span>-th generator.</li> <li><strong>acr relations:</strong> Minimal generating set of the ideal of relations of the anticanonical ring.</li> <li><strong>hsNum, hsDenom:</strong> Numerator and denominator of the Hilbert-Poincaré series, i.e. <span class="math-tex">\({\rm HP}_X(t) = {\rm hsNum}(t)/ {\rm hsDenom}(t)\)</span>.</li> <li><strong>hs0 - hs7:</strong> The <span class="math-tex">\(i\)</span>-th coefficient of the power series expansion of <span class="math-tex">\({\rm HP}_X(t)\)</span>, i.e. <span class="math-tex">\({\rm HP}_X(t) = {\rm hs0} + {\rm hs1}\, t + {\rm hs2}\, t^2 + \dots + {\rm hs7}\, t^7 + \dots\)</span></li> <li><strong>codim:</strong> The codimension of the embedding into a weighted projective space provided by the choice of generators of the anticanonical ring.</li> <li><strong>genus:</strong> This equals <span class="math-tex">\(h^0 (X, -\mathcal{K}_X)-2\)</span>.</li> </ul>
Temporal Distortion for Angry Faces: Testing Visual Attention and Action Preparation Accounts
<p>Datasets for Experiment 1 and Experiment 2 of "Temporal Distortion for Angry Faces: Testing Visual Attention and Action Preparation Accounts"</p>
Results of the crowd-mapping action within the project TeRRIFICA [Dataset No. 1 dated 2022-09-19]
<p>The dataset includes the results of the crowd-mapping action within the project "Territorial RRI fostering innovative climate action" - TeRRIFICA (Horizon 2020 under GA 824489) dated 2022-09-19. The data are points added to the map by the users (volunteers) and represent locations where climate change-related issues occur regarding air temperature, air quality, water, soil, and wind (SPOTS). The second part of the dataset is related to the crowd-mapping users and their anonymized characteristics (USERS). More details are available at https://terrifica.eu/.</p>
Mapping of Collective Action Opportunities for Water Stewardship
<p>With this global mapping exercise, we attempted to identify catchments with stronger need and potential of Collective Action for Water Stewardship. On the “need” side we considered a selection of water and biodiversity risk layers from the <a href="https://riskfilter.org/">WWF Risk Filter Suite</a>. On the “potential” side we considered economic factors such as value of crop production, density of business facilities (assets), and potential for cross-industry collaboration. The result is a global map (or a shapefile) of collective action opportunities, highlighting 350 catchments, across 100 river basins and 7 regions of the world, where multiple NGOs and the private sector shall work together to accelerate collective action for water stewardship. See the interactive map <a href="https://panda.maps.arcgis.com/apps/mapviewer/index.html?webmap=8d5c4bf6c6ff46c1a5c00e0b62e0b821">here</a>.</p> <p>This repository hosts all outputs as well as the input data of the mapping exercise. For transparency, reproducibility, as well as for future enhancements, the code is publicly available at <a href="https://github.com/rafaexx/collective_action_opportunities">https://github.com/rafaexx/collective_action_opportunities</a></p> <p>Final outputs are basically the shapefiles in <strong>output/summarise_geo/shp/*</strong> as well as the list of basins in <strong>output/summarise_geo/3_prioritization/hybas6_list_basins_collective_action.xlsx</strong></p> <p>For more details see part III of the report <a href="https://wwfint.awsassets.panda.org/downloads/unpacking-collective-action-in-water-stewardship.pdf">Unpacking Collective Action in Water Stewardship: Shared Solutions for Shared Water Challenges</a>.</p> <p> </p> <p> </p> <p> </p> <p> </p>
An integrated approach including docking, MD simulations, and network analysis highlights the action mechanism of the cardiac hERG activator RPR260243
<p>500 ns MD trajectories of the hERG bound states (protein and ligand) used for the analyses discussed in the paper. There are three replica for each system. The trajectories can be visualized using molecular visualization programs such as Pymol or VMD uploading the PDB and the DCD file.</p> <p>The PDB and the topology files of the hERG bound state (protein, membrane, ions and ligand) are also included.</p> <p>An example of input file used for the production step of the dynamics has been provided (production_1.conf). </p>
Content Analysis of Canada's Science Based Departments and Agencies Open Science Action Plans
<p>Dataset and codebook for a content analysis of Science-based departments and agencies open science action plans in response to the Government of Canada's Roadmap for Open Science. </p>
Green Intelligence Awareness, educational actions in Valladolid city
<p>Changes in behavior and human attitudes are fundamental to achieve a more sustainable world, so that, it is very interesting to analyze the potential of an activity or intervention to increase the green intelligence awareness of a population.</p> <p>There is enormous opportunity for nature-based solutions to promote understanding of sustainability in ways that positively influence citizen behavior. There are numerous available resources to learn and understand the fragility of our environmental and the responsibility of humans to protect, preserve and respect the world. Therefore, this KPI aims to reflect how the intervention is used for educational purposes and enhancement of public awareness. </p>
Keras video classification example with a subset of UCF101 - Action Recognition Data Set (top 10 videos)
<p>Classify video clips with natural scenes of actions performed by people visible in the videos.</p> <p>See the UCF101 Dataset web page: <a href="https://www.crcv.ucf.edu/data/UCF101.php#Results_on_UCF101">https://www.crcv.ucf.edu/data/UCF101.php#Results_on_UCF101</a></p> <p>This example datasets consists of the 10 most numerous video from the UCF101 dataset. For the top 5 version, see: <a href="https://doi.org/10.5281/zenodo.7924745">https://doi.org/10.5281/zenodo.7924745</a> .</p> <p>Based on this code: <a href="https://keras.io/examples/vision/video_classification/">https://keras.io/examples/vision/video_classification/</a> (needs to be updated, if has not yet been already; see the issue: <a href="https://github.com/keras-team/keras-io/issues/1342">https://github.com/keras-team/keras-io/issues/1342</a>).</p> <p>Testing if data can be downloaded from figshare with `wget`, see: <a href="https://github.com/mojaveazure/angsd-wrapper/issues/10">https://github.com/mojaveazure/angsd-wrapper/issues/10</a></p> <p>For generating the subset, see this notebook: <a href="https://colab.research.google.com/github/sayakpaul/Action-Recognition-in-TensorFlow/blob/main/Data_Preparation_UCF101.ipynb">https://colab.research.google.com/github/sayakpaul/Action-Recognition-in-TensorFlow/blob/main/Data_Preparation_UCF101.ipynb</a> -- however, it also needs to be adjusted (if has not yet been already - then I will post a link to the notebook here or elsewhere, e.g., in the corrected notebook with Keras example).</p> <p>I would like to thank Sayak Paul for contacting me about his example at Keras documentation being out of date. </p> <p>Cite this dataset as:</p> <p>Soomro, K., Zamir, A. R., & Shah, M. (2012). UCF101: A dataset of 101 human actions classes from videos in the wild. <em>arXiv preprint arXiv:1212.0402</em>. <a href="https://doi.org/10.48550/arXiv.1212.0402">https://doi.org/10.48550/arXiv.1212.0402</a></p> <p>To download the dataset via the command line, please use:</p> <pre><code class="language-bash">wget -q https://zenodo.org/record/7882861/files/ucf101_top10.tar.gz -O ucf101_top10.tar.gz tar xf ucf101_top10.tar.gz</code></pre> <p> </p>
Keras video classification example with a subset of UCF101 - Action Recognition Data Set (top 5 videos)
<p>Classify video clips with natural scenes of actions performed by people visible in the videos.</p> <p>See the UCF101 Dataset web page: <a href="https://www.crcv.ucf.edu/data/UCF101.php#Results_on_UCF101">https://www.crcv.ucf.edu/data/UCF101.php#Results_on_UCF101</a></p> <p>This example datasets consists of the 5 most numerous video from the UCF101 dataset. For the top 10 version see: <a href="https://doi.org/10.5281/zenodo.7882861">https://doi.org/10.5281/zenodo.7882861</a> .</p> <p>Based on this code: <a href="https://keras.io/examples/vision/video_classification/">https://keras.io/examples/vision/video_classification/</a> (needs to be updated, if has not yet been already; see the issue: <a href="https://github.com/keras-team/keras-io/issues/1342">https://github.com/keras-team/keras-io/issues/1342</a>).</p> <p>Testing if data can be downloaded from figshare with `wget`, see: <a href="https://github.com/mojaveazure/angsd-wrapper/issues/10">https://github.com/mojaveazure/angsd-wrapper/issues/10</a></p> <p>For generating the subset, see this notebook: <a href="https://colab.research.google.com/github/sayakpaul/Action-Recognition-in-TensorFlow/blob/main/Data_Preparation_UCF101.ipynb">https://colab.research.google.com/github/sayakpaul/Action-Recognition-in-TensorFlow/blob/main/Data_Preparation_UCF101.ipynb</a> -- however, it also needs to be adjusted (if has not yet been already - then I will post a link to the notebook here or elsewhere, e.g., in the corrected notebook with Keras example).</p> <p>I would like to thank Sayak Paul for contacting me about his example at Keras documentation being out of date. </p> <p>Cite this dataset as:</p> <p>Soomro, K., Zamir, A. R., & Shah, M. (2012). UCF101: A dataset of 101 human actions classes from videos in the wild. <em>arXiv preprint arXiv:1212.0402</em>. <a href="https://doi.org/10.48550/arXiv.1212.0402">https://doi.org/10.48550/arXiv.1212.0402</a></p> <p>To download the dataset via the command line, please use:</p> <pre><code class="language-bash">wget -q https://zenodo.org/record/7924745/files/ucf101_top5.tar.gz -O ucf101_top5.tar.gz tar xf ucf101_top5.tar.gz</code></pre>
Data Analysis files for "Coherent optical control of a superconducting microwave cavity via electro-optical dynamical back-action"
<p>Data analysis files for the manuscript "Coherent optical control of a superconducting microwave cavity via electro-optical dynamical back-action", <a href="https://www.nature.com/articles/s41467-023-39493-3#data-availability">Nature Communications <strong>14</strong>, 3784 (2023)</a>, or <a href="https://arxiv.org/abs/2210.12443">arXiv:2210.12443 (2022)</a></p> <p>This contains the raw data, the data analysis files, and the figure generation files of the manuscript, which includes the following three parts,</p> <p>0. Data preparation</p> <p>The total size of the raw dataset is around 120GB. The raw data is pre-processed via digital down-conversion at 40MHz to obtain the optical/microwave transient response of the electro-optical device in the presence of strong optical pulses at different powers and frequencies.</p> <p>The processed data is adopted for data analysis of the response measurements for convenience.</p> <p>1. Data Analysis</p> <ul> <li>Detailed data analysis of the electro-optical (microwave and optical) responses in presence of the optical pump pulses for different mode and probing configurations.</li> </ul> <p>2. Figures for the manuscripts.</p> <ol> <li>Figures for the optical characterizations</li> <li>Figures for the coherent responses for different configurations</li> <li>Figures for the excess back-action</li> <li>Figures for the theoretical curves in the Supplementary Information </li> </ol>
Audiovisual Moments in Time: A Large-Scale Annotated Dataset of Audiovisual Actions
<p>We present Audiovisual Moments in Time (AVMIT), a large-scale dataset of audiovisual action events. In an extensive annotation task 11 participants labelled a subset of 3-second audiovisual videos from the Moments in Time dataset (MIT). For each trial, participants assessed whether the labelled audiovisual action event was present and whether it was the most prominent feature of the video. The dataset includes the annotation of 57,177 audiovisual videos, each independently evaluated by 3 of 11 trained participants. From this initial collection, we created a curated test set of 16 distinct action classes, with 60 videos each (960 videos). We also offer 2 sets of pre-computed audiovisual feature embeddings, using VGGish/YamNet for audio data and VGG16/EfficientNetB0 for visual data, thereby lowering the barrier to entry for audiovisual DNN research. We further carried out an experiment to explore the utility of the AVMIT annotations and feature embeddings. A series of 6 Recurrent Neural Networks (RNNs) were trained on either AVMIT-filtered audiovisual events or modality-agnostic events from MIT, and then tested on our audiovisual test set. In all RNNs, top 1 accuracy was increased by 2.71-5.94\% by training exclusively on audiovisual events, even outweighing a three-fold increase in training data. We anticipate that the newly annotated AVMIT dataset will serve as a valuable resource for research and comparative experiments involving computational models and human participants, specifically when addressing research questions where audiovisual correspondence is of critical importance.</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.