A Survey on Methods for the Safety Assurance of Machine Learning Based Systems

Schwalbe, GesinaGesinaSchwalbe0000-0003-2690-2478Schels, MartinMartinSchels2020-03-112020-03-112020https://fis.uni-bamberg.de/handle/uniba/47326Methods for safety assurance suggested by the ISO 26262 automotive functional safety standard are not sufficient for applications based on machine learning (ML). We provide a structured, certification oriented overview on available methods supporting the safety argumen-tation of a ML based system. It is sorted into life-cycle phases, and maturity of the approach as well as applicability to different ML types are collected. From this we deduce current open challenges: powerful solvers, inclusion of expert knowledge, validation of data representativity and model diversity, and model introspection with provable guarantees.engfunctional safetylife-cycleISO 26262machine learningexplainable AI004A Survey on Methods for the Safety Assurance of Machine Learning Based Systemsconferenceobjecthttps://hal.archives-ouvertes.fr/hal-02442819v1