Thanh Bui

Markus Borg, Jens Henriksson, Kasper Socha et al.

Integration of Machine Learning (ML) components in critical applications introduces novel challenges for software certification and verification. New safety standards and technical guidelines are under development to support the safety of ML-based systems, e.g., ISO 21448 SOTIF for the automotive domain and the Assurance of Machine Learning for use in Autonomous Systems (AMLAS) framework. SOTIF and AMLAS provide high-level guidance but the details must be chiseled out for each specific case. We initiated a research project with the goal to demonstrate a complete safety case for an ML component in an open automotive system. This paper reports results from an industry-academia collaboration on safety assurance of SMIRK, an ML-based pedestrian automatic emergency braking demonstrator running in an industry-grade simulator. We demonstrate an application of AMLAS on SMIRK for a minimalistic operational design domain, i.e., we share a complete safety case for its integrated ML-based component. Finally, we report lessons learned and provide both SMIRK and the safety case under an open-source licence for the research community to reuse.

Thanh Bui, Tuomas Aura

There has recently been a flood of interest in potential new applications of blockchains, as well as proposals for more generic designs called public ledgers. Most of the novel proposals have been in the financial sector. However, the public ledger is an abstraction that solves several of the fundamental problems in the design of secure distributed systems: global time in the form of a strict linear order of past events, globally consistent and immutable view of the history, and enforcement of some application-specific safety properties. This paper investigates the applications of public ledgers to access control and, more specifically, to group management in distributed systems where entities are represented by their public keys and authorization is encoded into signed certificates. It is particularly difficult to handle negative information, such as revocation of certificates or group membership, in the distributed setting. The linear order of events and global consistency simplify these problems, but the enforcement of internal constraints in the ledger implementation often presents problems. We show that different types of revocation require slightly different properties from the ledger. We compare the requirements with Bitcoin, the best known blockchain, and describe an efficient ledger design for membership revocation that combines ideas from blockchains and from web-PKI monitoring. While we use certificate-based group-membership management as the case study, the same ideas can be applied more widely to rights revocation in distributed systems.