DatenVorsprung GmbH
Wien
Unternehmensphase:
Seed (Zumindest erste Testkunden)
Unternehmensform:
GmbH
Gegründet:
08.2020
MitarbeiterInnen:
9
Anbieter kontaktieren

DatenVorsprung GmbH

 bei aws Connect seit: 29.10.2024

Unternehmensprofil

"DatenVorsprung offers software solutions for testing and verification of AI systems."

DatenVorsprung GmbH is dedicated to enabling control over AI systems and removing risks in industrial AI applications. We develop formal verification, sensitivity analysis, robustness and certification solutions of AI applications across diverse industries, such as manufacturing, robotics, automotive, aerospace, marketing and finance. With a team of AI, mathematics and data science experts, we ensure that AI systems are effective, resilient, and secure for professional real-world applications.

Dieses Unternehmen ist in den folgenden Bereichen tätig::
Sensors, Artificial Intelligence, Robotics
Details:
Our patent-pending product TraceTube analyzes complex dynamical systems (AI or deterministic) for ensuring safety, performance and robustness. It observes system data from simulation or real world, and outputs a qualitative assessment for robustness and a quantitative risk probability. We ensure that AI systems work consistently under varying conditions and disturbances. We offer tailored AI solutions, verification integration, optimization of AI systems and legal compliance for certification.
Die KI-Lösungen wurden für folgende Branchen entwickelt:
Automotive & Luftfahrt, Energiewirtschaft, Finance & Insurance & Legal, Heathcare & BioTech, Landwirtschaft, Lifestyle & Entertainment, Mobilität & Transport, Produktion & Maschinenbau, Security & Big Data, Software & Technology, Tourismus & Gastronomie, Umwelt & Abfallwirtschaft
Die KI-Lösungen sind für folgende Anwendungsfälle relevant:
IT Automatisierung, Qualitätskontrolle , Predictive analysis, Vorhersagen, Entscheidungsunterstützung, Intelligente Geräte & Roboter, Risiko Management, Audit & Compliance
Folgende Technologien kommen zum Einsatz:
Computer Vision, Machine Learning, Neural Networks, Sonstige
Der KI-Anbieter kann sich in folgenden Bereichen eine Kooperation vorstellen:
Forschung & Entwicklung, Use Case / Pilotprojekt, Know-How Austausch, Organisatorische / IT-Ressourcen, Kunden- / Lieferanten-Beziehung, Produktion
Eine Schnittstelle (API) ist vorhanden:
Ja
Referenzen:
Beckhoff Automation GmbH
Data Scientist

Ein Data Scientist, der Sie bei Auswahl und Training von KI-Algorithmen unterstützen kann, ist vorhanden.

Data Engineer

Ein Data Engineer, der Sie bei Aufbau und Analyse einer strukturierten Datenbasis unterstützen kann, ist vorhanden.

Produkte / Dienstleistungen / Use Cases / Forschungsgebiete

AI Certification in Industrial Applications

TraceTube enables formal verification of machine learning algorithms for robustness and safety, aligning them with ISO 24029 certification requirements, which is essential for AI-driven industrial systems. By partnering with certification authorities (TÜV), Datenvorsprung accelerates AI certification and reduces costs. We provide safety and robustness tests that are integral to the certification process and help companies maintain legal compliance, building trust with customers and regulators.

Bridging the Simulation-to-Reality Gap in AI Systems

For the cart pole setup, TraceTube provided two insights for closing the Sim2Real gap: a definitive yes-or-no answer on whether simulated systems function safely in real-world conditions, and a quantitative understanding of the safe operational parameter ranges. TraceTube verifies that systems stay within safe limits when deployed. If the system stays within broad boundaries during simulation, we can have a more confident idea of the performance in the real world, which is less predictable.

Verification of an AI-controlled Cartpole System

In this cartpole setup, a magnetically driven cart on Beckhoff’s XTS rail must swing up and balance a pole using both sensor- and image-based controllers. TraceTube ensures system safety by verifying that the cart stays within rail boundaries during rapid accelerations and confirms stability during the transition from sensor- to video-based control, even with image-processing delays. TraceTube provides verification for safety-critical AI systems without requiring access to the model itself.