Machine Learning (ML) is increasingly used to drive the operation of complex distributed systems deployed on the cloud-edge continuum making their behavior non-deterministic. Their increasing adoption is urgently calling for assurance solutions assessing their non-functional properties (e.g., fairness, robustness, privacy) with the aim of improving trustworthiness. Certification has been clearly identified by policymakers, regulators, and industrial stakeholders as the reliable assurance solution to address this pressing need. Unfortunately, existing certification schemes are not immediately applicable to systems whose non-deterministic behavior is built on ML models. This article analyzes the challenges and deficiencies of current certification schemes, discusses open research issues, and proposes a first certification scheme for ML-based distributed systems behavior.
翻译:暂无翻译