Skip to main content
SpringerLink
Log in

CertML: Initial Steps Towards Using N-Version Neural Networks for Improving AI Safety

  • Schwerpunkt
  • Published:
Datenschutz und Datensicherheit - DuD Aims and scope Submit manuscript

Zusammenfassung

Currently, neural networks can only be used to a limited extent in safety-critical applications, because their quality and safety properties are difficult or impossible to prove due to their inherent complexity and their black-box character. The project CertML aims to improve and verify the safety of machine learning based systems.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

References

  1. Bilanz zur Verkehrssicherheitslage 2021in Berlin (Feb 2022). URL: https://www.berlin.de/polizei/polizeimeldungen/2022/pressemitteilung.1178821.php

  2. Abbiegeassistenten ab sofort Pflicht, Bundesregierung (Jul 2022). URL: https://www.bundesregierung.de/breg-de/themen/europa/ abbiegeassistenten-pflicht-2060086

  3. F. Machida: N-version machine learning models for safety critical systems, In: 2019, 49th Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshops (DSNW), 2019, pp. 48–51. https://doi.org/10.1109/DSN-W.2019.00017.

  4. T. Michalke, C. Wüst, D. Feng, M. Dolgov, C. Gläser, F. Timm: Where can i drive? a system approach: Deep ego corridor estimation for robust automated driving, In: 2021 IEEE International Intelligent Transportation Systems Conference (ITSC), IEEE, 2021, pp. 1565–1571.

  5. D. Gupta: Image segmentation keras: Implementation of segnet, fcn, unet, pspnet and other models in keras, https://github.com/divamgupta/image-segmentation-keras (2022).

Download references

Author information

Authors and Affiliations

  1. HTW Berlin, Berlin, Deutschland

    Philipp Jass, Hamzih Abukhashab & Carsten Thomas

  2. 3DResearch Fahrzeugelektronik GmbH, Berlin, Deutschland

    Peter Woltersdorf & Michael Weber

  3. samoconsult GmbH, Berlin, Deutschland

    Mirko Conrad, Ines Fey & Harald Schülzke

Authors
  1. Philipp Jass
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Hamzih Abukhashab
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Carsten Thomas
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Peter Woltersdorf
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Michael Weber
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Mirko Conrad
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Ines Fey
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Harald Schülzke
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Philipp Jass.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Jass, P., Abukhashab, H., Thomas, C. et al. CertML: Initial Steps Towards Using N-Version Neural Networks for Improving AI Safety. Datenschutz Datensich 47, 483–486 (2023). https://doi.org/10.1007/s11623-023-1803-z

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11623-023-1803-z

Search

Navigation