No abstract

Emission cleaning in modern cars is controlled by embedded software. In this context, the diesel emission scandal has made it apparent that the automotive industry is susceptible to fraudulent behaviour, implemented and effectuated by that control software. Mass effects make the individual controllers altogether have statistically significant adverse effects on people’s health. This paper surveys recent work on the use of rigorous formal techniques to attack this problem. It starts off with an introduction into the dimension and facets of the problem from a software technology perspective. It then details approaches to use (i) model checking for the white-box analysis of the embedded software, (ii) model- based black-box testing to detect fraudulent behaviour under standardized conditions, and (iii) synthesis of runtime monitors for real driving emissions of cars in-the-wild. All these efforts aim at finding ways to eventually ban the problem of doped software, that is, of software that surreptitiously alters its behaviour in certain circumstances – against the interest of the owner or of society.

No abstract

We argue that, to be trustworthy, Computational Intelligence (CI) has to do what it is entrusted to do for permissible reasons and to be able to give rationalizing explanations of its behavior which are accurate and graspable. We support this claim by drawing parallels with trustworthy human persons, and we show what difference this makes in a hypothetical CI hiring system. Finally, we point out two challenges for trustworthy CI and sketch a mechanism which could be used to generate sufficiently accurate as well as graspable rationalizing explanations for CI behavior.

Future cyber-physical systems are expected to be dynamic, evolving while already being deployed. Frequent updates of software components are likely to become the norm even for safety-critical systems. In this setting, a full re-certification before each software update might delay important updates that fix previous bugs, or security or safety issues. Here we propose a vision addressing this challenge, namely through the evidence-based continuous supervision and certification of software variants in the field. The idea is to run both old and new variants of component software inside the same system, together with a supervising instance that monitors their behavior. Updated variants are phased into operation after sufficient evidence for correct behavior has been collected. The variants are required to explicate their decisions in a logical language, enabling the supervisor to reason about these decisions and to identify inconsistencies. To resolve contradictory information, the supervisor can run a component analysis to identify potentially faulty components on the basis of previously observed behavior, and can trigger micro-experiments which plan and execute system behavior specifically aimed at reducing uncertainty. We spell out our overall vision, and provide a first formalization of the different components and their interplay. In order to provide efficient supervisor reasoning as well as automatic verification of supervisor properties we introduce SupERLog, a logic specifically designed to this end.