The AI-in-the-Loop toolchain facilitates Continuous Improvement and Continuous Development (CI/CD) of artificial intelligence (AI), with a strong focus on ensuring the safety and reliability of AI applications.
AI-driven DevOps and Verification & Validation
Continuous Delivery in the AI-in-the-loop toolchain for the development of safe and reliable AI applications.
Credit:
DLR, IAV, Hochschule Mittweida
Continuous Development and Improvement of AI
Co-developed with IAV, this toolchain supports the real-time training of AI algorithms in real-world settings, providing immediate insights into AI behavior. As a laboratory for automated service robots, we enable real-time monitoring to capture direct feedback on AI behavior and algorithm adaptation during the training process.
Consequently, this allows for continuous feedback on performance and adjustments throughout the training phase. This approach substantially accelerates the optimization of safety-critical AI systems, offering faster and more efficient improvements compared to offline methods.
As part of the "AI-in-the-Loop" project, a solution has been developed that enables the seamless integration of deployed service robots into a data-service ecosystem. Through this ecosystem, the robots can not only be efficiently configured but also flexibly adapted to different use cases.
