Replication Data for: "DE-VAE: Revealing Uncertainty in Parametric and Inverse Projections with Variational Autoencoders using Differential Entropy"

Version 1.0

Dennig, Frederik L.; Keim, Daniel, 2026, "Replication Data for: "DE-VAE: Revealing Uncertainty in Parametric and Inverse Projections with Variational Autoencoders using Differential Entropy"", https://doi.org/10.18419/DARUS-5258, DaRUS, V1

Learn about Data Citation Standards.

Contact Owner

Dataset Metrics

8 Downloads

Description	This is the replication data for our paper "DE-VAE: Revealing Uncertainty in Parametric and Inverse Projections with Variational Autoencoders using Differential Entropy", allowing additional analysis and experiments. The files, per file, below. It includes the experiment data and source code. For details on the usage, please refer to the README.md in the "source-code.zip" file. All instructions on how to use the method can be found there. (2026-02-18)
Subject	Computer and Information Science
Keyword	Uncertainty, Dimensionality Reduction, Autoencoder, Data Visualization
Related Publication	Is Cited By: F. L. Dennig and D. A. Keim, "DE-VAE: Revealing Uncertainty in Parametric and Inverse Projections with Variational Autoencoders using Differential Entropy," in 2025 IEEE Workshop on Uncertainty Visualization: Unraveling Relationships of Uncertainty, AI, and Decision-Making, Vienna, Austria, 2025, pp. 33-37.doi: 10.1109/UncertaintyVisualization68947.2025.00009
Notes	Use persistent identifiers from Software Heritage ( ) to cite individual files or even lines of the source code.
License/Data Use Agreement	Custom Dataset Terms

Filter by

	1 to 4 of 4 Files	Download
	models.tar.xz Compressed Archive - 2.2 GB Published Feb 20, 2026 2 Downloads MD5: 66da0557ed92a25f06d148d95ea87a2a All the trained models from the experiments. All models can be loaded using the functionality provided by our program available in source-code.zip. The file names indicate the model type and parameters.	Access File File Access Public Download Options Compressed Archive Download Metadata Data File Citation Download EndNote XML Download RIS Download BibTeX
	preprocessed.tar.xz Compressed Archive - 24.0 MB Published Feb 20, 2026 2 Downloads MD5: 598dfe09e99ada611261ad6d51dafb81 This includes all pre-processed datasets from our experiments, including projected datasets, CSV files, and serialised NumPy arrays.	Access File File Access Public Download Options Compressed Archive Download Metadata Data File Citation Download EndNote XML Download RIS Download BibTeX
	records.tar.xz Compressed Archive - 828.9 KB Published Feb 20, 2026 2 Downloads MD5: 8ebc1835aae7854c3973a3a5b05dc7db All records gathered during training are included, such as training and validation loss, as well as the individual loss components.	Access File File Access Public Download Options Compressed Archive Download Metadata Data File Citation Download EndNote XML Download RIS Download BibTeX
	source-code.zip ZIP Archive - 22.0 MB Published Feb 20, 2026 2 Downloads MD5: 87a5dbba03be2d0f7f88a5eb0edc6047 The source code of our approach. It trains neural networks and logs the results.	Preview "source-code.zip" Access File File Access Public Download Options ZIP Archive Download Metadata Data File Citation Download EndNote XML Download RIS Download BibTeX

Citation Metadata

Persistent Identifier	doi:10.18419/DARUS-5258
Publication Date	2026-02-20
Title	Replication Data for: "DE-VAE: Revealing Uncertainty in Parametric and Inverse Projections with Variational Autoencoders using Differential Entropy"
Alternative URL	https://osf.io/zr6xf
Other Identifier	Software Heritage: https://archive.softwareheritage.org/swh:1:dir:510a288c701dcf5d34559584c7ce5f80f7a55b61;origin=https://github.com/fredooo/DE-VAE;visit=swh:1:snp:ec8008a13092d0c361223ad9578f692c09b1a490;anchor=swh:1:rev:58d41486aba77913b892a6e4b9bb5e9c7796e031
Author	https://ror.org/0546hnb39https://orcid.org/0000-0003-1116-8450 https://ror.org/0546hnb39https://orcid.org/0000-0001-7966-9740
Point of Contact	Use email button above to contact. Keim, Daniel (University of Konstanz)
Description	This is the replication data for our paper "DE-VAE: Revealing Uncertainty in Parametric and Inverse Projections with Variational Autoencoders using Differential Entropy", allowing additional analysis and experiments. The files, per file, below. It includes the experiment data and source code. For details on the usage, please refer to the README.md in the "source-code.zip" file. All instructions on how to use the method can be found there. (2026-02-18)
Subject	Computer and Information Science
Keyword	Uncertainty https://www.ncbi.nlm.nih.gov/mesh/68035501 (MeSH) https://www.ncbi.nlm.nih.gov/mesh/ Dimensionality Reduction https://www.ncbi.nlm.nih.gov/mesh/2108071 (MeSH) https://www.ncbi.nlm.nih.gov/mesh/ Autoencoder https://www.ncbi.nlm.nih.gov/mesh/2108103 (MeSH) https://www.ncbi.nlm.nih.gov/mesh/ Data Visualization https://www.ncbi.nlm.nih.gov/mesh/2028190 (MeSH) https://www.ncbi.nlm.nih.gov/mesh/
Topic Classification	Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing (DFGFO) https://w3id.org/dfgfo/2024/443-05 Artificial Intelligence and Machine Learning Methods (DFGFO) https://w3id.org/dfgfo/2024/443-04
Related Publication	Is Cited By: F. L. Dennig and D. A. Keim, "DE-VAE: Revealing Uncertainty in Parametric and Inverse Projections with Variational Autoencoders using Differential Entropy," in 2025 IEEE Workshop on Uncertainty Visualization: Unraveling Relationships of Uncertainty, AI, and Decision-Making, Vienna, Austria, 2025, pp. 33-37. doi 10.1109/UncertaintyVisualization68947.2025.00009 https://doi.org/10.1109/UncertaintyVisualization68947.2025.00009
Notes	Use persistent identifiers from Software Heritage ( ) to cite individual files or even lines of the source code.
Producer	Dennig, Frederik L. (University of Konstanz)
Project	SFB/Transregio 161
Funding Information	DFG: 251654672
Distributor	Dennig, Frederik L. (University of Konstanz)
Depositor	Dennig, Frederik L.
Deposit Date	2026-02-18
Software	Python, Version: 3.11 PyIP, Version: 25.2 virtualenv, Version: 20.32
Data Source	Blobs: D. Blumberg, Y. Wang, A. Telea, D. A. Keim, and F. L. Dennig, "Inverting Multidimensional Scaling Projections Using Data Point Multilateration," in 15th International EuroVis Workshop on Visual Analytics, 2024.; HAR: D. Anguita, A. Ghio, L. Oneto, X. Parra, and J. L. Reyes-Ortiz, "Human activity recognition on smartphones using a multiclass hardware-friendly support vector machine," in Ambient Assisted Living and Home Care, 2012, pp. 216–223.; MNIST: Y. LeCun, C. Cortes, and C. Burges, "MNIST handwritten digit database," 1998.; Fashion-MNIST: H. Xiao, K. Rasul, and R. Vollgraf, "Fashion-MNIST: A Novel Image Dataset for Benchmarking Machine Learning Algorithms," CoRR, vol. abs/1708.07747, 2017.; KMNIST: T. Clanuwat, M. Bober-Irizar, A. Kitamoto, A. Lamb, K. Yamamoto, and D. Ha, "Deep Learning for Classical Japanese Literature," CoRR, vol. abs/1812.01718, 2018.
Documentation and Access to Sources	The HAR dataset is part of the source code; all other datasets are available via the torchvision library imported by the project. All are well-known datasets for machine learning problems.
Did it work?	Yes
Explanation	We were able to model uncertainty through three distinct latent Gaussian structures. However, quantitative performance is inconsistent across datasets and projections. The loss-weight parameterization remains dataset-dependent and challenging.

Process Metadata

Processing Methods

Network Training (Setup (per run): seed is fixed for reproducibility, data loaders are created (80/10/10 train/val/test split), model and Adam optimizer (lr=1e-3) are initialized.
Per-epoch loop (up to 100 epochs):
1. Training step - model set to train mode; for each batch of (input vector, 2D projection point, label), a forward pass produces a reconstruction, latent mean, and covariance parameter; loss is computed as reconstruction + weighted projection - weighted entropy; gradients are back-propagated and weights updated.
2. Validation step - same loss computed on the validation set without gradients.
3. Early stopping (patience=5) - if validation loss improved, model is saved to disk; if no improvement for 5 consecutive epochs, training halts.
Loss components (total, reconstruction, projection, entropy, avg std) are logged to disk per epoch for both train and val splits.
Post-training: Final loss evaluated on the test set and logged.
Full experiment sweep: All combinations of 4 model types x datasets x projections x 10 seeds are trained sequentially, with dataset/projection-specific loss weights.)

Software

Python Version: 3.11 (https://www.python.org/) Python Software Foundation License (PSFL)
PyIP Version: 25.2 (https://pypi.org/) Python Software Foundation License (PSFL)
virtualenv Version: 20.32 (https://virtualenv.pypa.io/en/latest/index.html) MIT License

Software Metadata (CodeMeta v2.0)

Software Version	v1.0.0
Development Status	Active
Code Repository	https://github.com/fredooo/DE-VAE
Application Category	Analysis and Visualization
Application Subcategory	Data Projection
Programming Language	Python
Runtime Platform	Python 3.11
Build Instructions	https://raw.githubusercontent.com/fredooo/DE-VAE/refs/heads/main/README.md
Software Requirements	Python 3.11 https://www.python.org/ PyIP 25.2 https://pip.pypa.io/en/stable/ virtualenv 20.32 https://virtualenv.pypa.io/en/latest/index.html
Readme	https://raw.githubusercontent.com/fredooo/DE-VAE/refs/heads/main/README.md
Release Notes	https://github.com/fredooo/DE-VAE/releases/tag/v1.0.0

Privacy Metadata

Personal Data	no

Dataset Terms

License/Data Use Agreement

Our Community Norms as well as good scientific practices expect that proper credit is given via citation. Please use the data citation shown on the dataset page.

Custom Dataset Terms - the following Custom Dataset Terms have been defined for this dataset.

The files models.tar.xz, preprocessing.tar.xz, and records.tar.xz are licensed under CC-By 4.0 (see also: https://osf.io/zr6xf) File source-code.zip is licensed under the MIT License (see also: https://github.com/fredooo/DE-VAE)

Dataset Version	Summary	Version Note	Edited by	Published on
No records found.

Edit File

This file has already been deleted (or replaced) in the current version. It may not be edited.

Restrict Access

Restricting limits access to published files. People who want to use the restricted files can request access by default. If you disable request access, you must add information about access to the Terms of Access field.

Learn about restricting files and dataset access in the User Guide.

Request Access

Enable access request

You must enable request access or add terms of access to restrict file access.

Terms of Access for Restricted Files

Save Changes

Edit Embargo

The selected file or files have already been published. Contact an administrator to change the embargo date or reason of the file or files.

Edit Retention Period

The selected file or files have already been published. Contact an administrator to change the retention period date or reason of the file or files.

Delete Files

The file will be deleted after you click on the Delete button.

Files will not be removed from previously published versions of the dataset.

Select File(s)

Please select one or more files.

Share Dataset

Share this dataset on your favorite social media networks.

Continue

Dataset Citations

Citations for this dataset are retrieved from Crossref via DataCite using Make Data Count standards. For more information about dataset metrics, please refer to the User Guide.

Sorry, no citations were found.

Inaccessible Files Selected

The selected file(s) may not be downloaded because you have not been granted access or the file(s) have a retention period that has expired or the files can only be transferred via Globus.

Ineligible Files Selected

The selected file(s) may not be transferred because you have not been granted access or the file(s) have a retention period that has expired or the files are not Globus accessible.

Download Options

The files selected are too large to download as a ZIP.

You can select individual files that are below the 1.9 GB download limit from the files table, or use the Data Access API for programmatic access to the files.

Select File(s)

Please select a file or files to be downloaded.

Inaccessible Files Selected

The selected file(s) may not be downloaded because you have not been granted access or the file(s) have a retention period that has expired.

Click Continue to download the files you have access to download.

Ineligible Files Selected

Some file(s) cannot be transferred. (They are restricted, embargoed, with an expired retention period, or not Globus accessible.)

Click Continue to transfer the elligible files.

Delete Dataset

Are you sure you want to delete this dataset and all of its files? You cannot undelete this dataset.

Delete Draft Version

Are you sure you want to delete this draft version? Files will be reverted to the most recently published version. You cannot undelete this draft.

Unpublished Dataset Preview URL

Preview URL can only be used with unpublished versions of datasets.

Unpublished Dataset Preview URL

Are you sure you want to disable the Preview URL? If you have shared the Preview URL with others they will no longer be able to use it to access your unpublished dataset.

Delete Files

The file(s) will be deleted after you click on the Delete button.

Files will not be removed from previously published versions of the dataset.

Compute

This dataset contains restricted files you may not compute on because you have not been granted access.

Deaccession Dataset

Are you sure you want to deaccession? This is permanent and the selected version(s) will no longer be viewable by the public.

Deaccession Dataset

Are you sure you want to deaccession this dataset? This is permanent, and it will no longer be viewable by the public.

Version Differences Details

Please select two versions to view the differences.

Version Differences Details

Version:
Last Updated:

Version Note:

Version:
Last Updated:

Version Note:

Select File(s)

Please select a file or files for access request.

Select File(s)

Embargoed files cannot be accessed. Please select an unembargoed file or files for your access request.

Edit Tags

Select existing file tags or create new tags to describe your files. Each file can have more than one tag.

Request Access

You need to Log In to request access.

Dataset Terms

Please confirm and/or complete the information needed below in order to request access to files in this dataset.

This dataset is made available under the following terms. Please confirm and/or complete the information needed below in order to continue.

License/Data Use Agreement

Our Community Norms as well as good scientific practices expect that proper credit is given via citation. Please use the data citation shown on the dataset page.

Custom terms specific to this dataset Custom Dataset Terms - the following Custom Dataset Terms have been defined for this dataset.

Terms of UseThe files models.tar.xz, preprocessing.tar.xz, and records.tar.xz are licensed under CC-By 4.0 (see also: https://osf.io/zr6xf) File source-code.zip is licensed under the MIT License (see also: https://github.com/fredooo/DE-VAE)

Preview Guestbook

Upon downloading files the guestbook asks for the following information.

Guestbook Name

Collected Data

Account Information

Package File Download

Use the Download URL in a Wget command or a download manager to download this package file. Download via web browser is not recommended. User Guide - Downloading a Dataverse Package via URL

Download URL

https://darus.uni-stuttgart.de/api/access/datafile/

Compute Batch

Clear Batch

Dataset	Persistent Identifier	Change Compute Batch

Compute Batch

Submit for Review

Please use the publication checklist for authors ( https://www.izus.uni-stuttgart.de/en/fokus/darus/publication/ ) to ensure publication readiness. You will not be able to make changes to this dataset while it is in review.

Please be advised that by clicking on Submit for Review a notification will be sent to your institution's local curator. The DaRUS team will NOT be notified unless the local curator or anyone with publishing rights starts the Publication workflow

Publish Dataset

Are you sure you want to republish this dataset?

Select if this is a minor or major version update.

Minor Release (1.1)

Major Release (2.0)

Version Note

Publish Dataset

This dataset cannot be published until SFB-TRR 161 A03 "Quantification of Visual Analytics Transformations and Mappings" is published by its administrator.

Publish Dataset

This dataset cannot be published until SFB-TRR 161 A03 "Quantification of Visual Analytics Transformations and Mappings" and SFB/Transregio 161 "Quantitative Methods for Visual Computing" are published.

Return to Author

Return this dataset to contributor for modification. The reason for return entered below will be sent by email to the author.

Curation Status History

Status	Date	Assigner
	2026-02-20 12:54:00	Dimitar Garkov
Waiting for Feedback from Author	2026-02-20 07:03:01	Dorothea Iglezakis
In Review by FoKUS	2026-02-18 15:47:22	Dorothea Iglezakis

Add/Edit a Version Note

Enter the reason this version was created. To learn more about Version Notes, visit the Version Notes section of the User Guide.

Version Note

Styled Citation