P1023: Prüfkriterien für Reinforcement Learning (RL-Criteria)

TENDER DESCRIPTION

The project develops a catalogue of requirements for assessing the security and trustworthiness of AI applications that use Reinforcement Learning. The requirements aim to cover the widest possible spectrum of applications. To this end, all techniques and variants of Reinforcement Learning are fundamentally taken into account, in particular also in connection with large and current generative models (keywords: RLHF, RLAIF). In the specific development, the requirements are supplemented in each case by explanations, accompanying information, implementation instructions (e.g. explanations regarding the use of suitable test methods), and similar content. The catalogue thus provides a basis for self-assessment and for testing relevant AI systems. The requirements developed in the project are prepared in such a way that they are compatible with the BSI’s SdT library and fit into the existing BSI product portfolio for AI-specific test criteria. Specifically, the requirements are to be designed in the end as a supplementary module to BSI A5. The project develops a catalogue of requirements for assessing the security and trustworthiness of AI applications that use Reinforcement Learning.…

TENDER BRIEF

TENDER DOCUMENTS