The 5th edition of the international workshop on Formal Approaches for Advanced Computing Systems (FAACS 2021) is co-located with the 15th European Conference on Software Architecture (ECSA). This year the workshop will be focusing on formal approaches for AI-based systems. It will be held virtually September TBD, 2021.
Advanced computing systems are increasingly adopted to generate productivity gains, improve well-being and help experts to address challenges in various application domains, including mobility, healthcare, cyber-security, industrial robots, autonomous vehicles, cloud-based services, and IoT. Such systems often operate in dynamic, highly complex, partially observable environments and they are affected by external uncertainty as well as internal uncertainty, very often due to “intelligent” and “learning” components that are becoming established and effective for making predictions, recommendations or decisions that influence the surroundings. As a consequence, the design and development of these systems require proper approaches, methodologies, abstractions and architectural design solutions integrating together heterogeneous elements such as networking, computation, engineered physical elements, AI/ML components, and human beings into complex ecosystems. Assuring dependability, trustworthiness and other important qualities of such systems yield challenges that call for novel and formal approaches to system design, development, validation, and verification. Indeed, an unexpected exposure of users to unwanted behaviors, such as threats to confidentiality or even financial or health danger, is highly undesirable.
The goal of the workshop is to foster integration between formal methods and software architectures, which allows for promoting new connections and synergies between the two research communities. An integrated research will help them address the challenges of the upcoming generation of advanced computing systems in the modern digital society.
We especially encourage contributions describing application of formal approaches to engineering AI-based systems.
Besides the special interest of this edition, the call is open to all the areas reported below and applied to engineer advanced systems in different contexts and domains (e.g., IoT, robotic systems, cyber-physical systems, autonomic computing, service-based systems, reconfigurable and self-adaptive systems, mobile and cloud applications). Areas of interest include, but are not limited to:
Submissions must follow the two-column CEUR-ART style (download template). All papers should be submitted before the submission deadline (see below) using the online submission site: EasyChair FAACS 2021 Workshop.
We solicit the following contribution types:
ECSA 2021 will use a two-step process for workshop proceedings. Online proceedings (available before the start of the conference) will include all the accepted papers (5 to 10 pages) of the workshops and will be published at CEUR (http://ceur-ws.org). After the conference, we will organise post-proceedings of selected and extended papers of workshops that will be published in a Springer LNCS volume (16 to 18 pages). Workshop papers submitted for the post-proceedings will undergo an additional review cycle.
September 14, 2021 virtual event
Joint program with MDE4SA:
Bio: Paolo Arcaini is project associate professor at the National Institute of Informatics (NII), Japan. He received a PhD in Computer Science from the University of Milan, Italy, in 2013. Before joining NII, he held an assistant professor position at Charles University, Czech Republic. His main research interests are related to search-based testing, fault-based testing, model-based testing, and automatic repair. His current research in the ERATO Metamathematics for Systems Design (MMSD) project is related to the testing and debugging of autonomous driving systems. More info at https://group-mmm.org/~arcaini/
Title:Exploiting system parameters to test and debug Autonomous Driving Systems
Abstract: Autonomous Driving Systems (ADSs) are safety-critical systems, whose testing is of paramount importance. However, ADS testing raises several challenges that are specific of the domain. Indeed, typical testing activities performed for software programs are not directly applicable to ADSs; first, defining a sufficiency testing criterion that targets all the different types of ADS behaviours is challenging; moreover, there is no proper oracle that can always specify the desired, correct behaviour. In this talk, we review a series of works that we proposed to target these challenges for a particular ADS, the path planner provided by our industry partner. The path planner decides which path to follow through a cost function that uses parameters to assign a cost to the driving characteristics (e.g., lateral acceleration or speed) that must be applied in the path. These parameters implicitly describe the behaviour of the path planner. We present how we have exploited this idea for defining a coverage criterion and a corresponding test generator, for automatically specifying an oracle, and for doing fault localization.