AI-powered Digital Assistant (AIDA)

The AI-powered Digital Assistant (AIDA) was developed under ESA TDE activity 1000033025 to explore use cases and technologies able to assist systems engineers. AIDA is able to analyse systems engineering models and:

• UC1: propose requirement flow-down, traceability between existing requirements and evaluation of their verification method
• UC2: propose new elements (functions/components) for existing system architectures
• UC3: check the coherency between requirements and architectures

For this, different machine learning techniques are used including Natural Language Processing (NLP), classification, clustering and Large Language Models (LLMs).

The estimated Technology Readiness Level (TRL) is 4.

The released data pack contains:

• A series of documents describing the use cases, technologies used, validation, etc. as well as the software user manual
• The source and application code: trained models, training code and dummy datasets for training

Please, log in in order to see the download links.

You can direct your comments or questions to alberto.gonzalez.fernandez@ext.esa.int.