Description

AI systems are increasingly present in everyday society, from simple computer systems to agents such as autonomous vehicles or social robots. In this context, it is often noted that it is critical to understand how human users perceive such systems - in particular, the degree to which they understand how the system works, and what mental models they build of the underlying algorithms. "Explainable AI" thus refers to AI systems that behave or provide the necessary information so that their working becomes comprehensible to the human user. The overall term can, however, cover various sub-disciplines ranging from machine learning researchers seeking to understand the functionality of black-box networks over researchers in decision support systems working on system awareness to cognitive scientists seeking to understand how humans understand intentions and other mental states in other agents.

List of Topics

The purpose of the present symposium is to bring together researchers from all aspects of XAI, and to foster an exchange of the current state of the art while facilitating the development of connections between different subfields of XAI. This is therefore intentionally intended to be broad so as to facilitate a diverse symposium that can cover diverse facets of XAI. Example topics include (but are certainly not restricted to):

• Human expectations of intelligent systems
• Methods by which various types of intelligent systems can generate explanations that are in line with human expectations.
• XAI systems’ design issues.
• Evaluation of the effectiveness of such systems, methods and metrics.
• Theoretical aspects of XAI. Definitions, challenges and future research.

Submission Guidelines

We will consider both extended abstracts (up to two pages) for poster presentation, and full papers (up to 8 pages) for oral or poster presentation.

Please use the AISB template and submit a folder with the full source files (including images) if you use Latex, or both a Word and a PDF version if you prefer to use the Word template. Submissions are via EasyChair.

Submission deadline: 15 February 2019
Notification of acceptance: 8 March 2019
Camera-ready submission: 25 March 2019

Invited Speakers

• Tarek Besold, AI Lead & Senior Research Scientist "Trustworthy AI" @ Alpha Health AI Lab. Barcelona, Catalonia, Spain.

Organizing Committee

• Serge Thill, Donders Institute for Brain, Cognition, and Behaviour, Netherlands
• Maria Riveiro, University of Skövde, Sweden

Contact

All questions about submissions should be emailed to the organizers at s.thill@donders.ru.nl or maria.riveiro@his.se