Implementation of airborne ML models with semantics preservation

2025-09-23

AI Summary

Airborne machine learning (ML) systems face critical challenges in runtime reliability, performance retention under target operational conditions, and semantic consistency between models and their specifications—under stringent aviation safety constraints. Method: This paper proposes a Machine Learning Model Description (MLMD) framework that rigorously separates executable ML models from their formal semantic descriptions. Integrating formal modeling, semantic equivalence verification, and certifiable model transformation techniques, the framework adheres strictly to EASA Concept Paper and ED-324 airworthiness standards to establish an assurance-driven, certifiable airborne ML development lifecycle. Contribution/Results: It introduces, for the first time, a systematic semantic preservation mechanism embedded across the entire ML lifecycle—guaranteeing strict behavioral equivalence between deployed models and their training-time specifications. Empirical evaluation across multiple industrial use cases demonstrates the framework’s feasibility and effectiveness in achieving both semantic fidelity and performance stability in realistic avionic environments.

Technology Category

• Machine Learning: Hardware-aware ML
• Humans and AI: Human-in-the-loop Machine Learning
• Application Domains: Mobility, Driving & Flight

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Security and Privacy: Security and privacy of machine learning and AI applications

Abstract

Machine Learning (ML) may offer new capabilities in airborne systems. However, as any piece of airborne systems, ML-based systems will be required to guarantee their safe operation. Thus, their development will have to be demonstrated to be compliant with the adequate guidance. So far, the European Union Aviation Safety Agency (EASA) has published a concept paper and an EUROCAE/SAE group is preparing ED-324. Both approaches delineate high-level objectives to confirm the ML model achieves its intended function and maintains training performance in the target environment. The paper aims to clarify the difference between an ML model and its corresponding unambiguous description, referred to as the Machine Learning Model Description (MLMD). It then refines the essential notion of semantics preservation to ensure the accurate replication of the model. We apply our contributions to several industrial use cases to build and compare several target models.

Problem

• Ensuring safety compliance of ML models in airborne systems
• Clarifying differences between ML models and their unambiguous descriptions
• Refining semantics preservation for accurate ML model replication

Innovation

• ML model unambiguous description for semantics preservation
• Refining semantics preservation for accurate model replication
• Applying contributions to industrial use cases comparison