Banner
Home      Log In      Contacts      FAQs      INSTICC Portal
 
Documents
Special Session
Special Session on
Interpretable Artificial Intelligence Through Glass-Box Models - IAI 2025

23 - 25 February, 2025 - Porto, Portugal

Within the 17th International Conference on Agents and Artificial Intelligence - ICAART 2025


CHAIR

Mattias Wahde
Department of Applied Mechanics, Chalmers University of Technology
Sweden

 
Brief Bio
Mattias Wahde is professor of Applied Artificial Intelligence at Chalmers University of Technology in Göteborg, Sweden. His research interests include natural language processing and conversational agents, stochastic optimization methods, robotics, and automated driving. The unifying theme of his research is the development and use of transparent, accountable, and safe methods (glass-box models) in artificial intelligence. He is the author or co-author of numerous papers in these fields, as well as a course book on stochastic optimization. In addition to research and PhD student supervision, he is heavily involved in teaching, as lecturer and examiner in several courses with a total of around 300 students per year. He is currently leading a theme on Interpretable AI in the framework of Chalmers' AI Research Centre (CHAIR).

SCOPE

This session is centered on AI models that prioritize transparency and human interpretability, commonly known as glass-box models. These models are essential in scenarios where accountability, safety, and ethical considerations are primary concerns. While black-box models (like deep neural networks, including large language models) achieve high performance, they lack transparency, making them prone to bias and adversarial vulnerabilities. By contrast, glass-box models (e.g., decision trees, rule-based systems, etc.) offer interpretability but sometimes at the cost of reduced performance. This session seeks papers that involve innovative approaches to enhance the performance and applicability of glass-box models. We also welcome contributions that explore the accuracy-interpretability trade-off, including theoretical analyses, empirical studies, and advancements that aim to balance these competing demands.

TOPICS OF INTEREST

Topics of interest include, but are not limited to:
  • Accuracy vs. interpretability tradeoff
  • Advancements in glass-box models
  • Ethical and social implications
  • Future directions in accountable AI
  • High-stakes applications (e.g., automated driving, healthcare, personal robotics)
  • Model complexity and interpretability
  • Model interpretability techniques
  • Robustness and security
  • Transparency in decision-making
  • User trust and adoption

IMPORTANT DATES

Paper Submission: December 23, 2024
Authors Notification: January 14, 2025
Camera Ready and Registration: January 22, 2025

SPECIAL SESSION PROGRAM COMMITTEE

Mauro Bellone, Tallinn University of Technology, Estonia
Marco L. Della Vedova, Chalmers University of Technology, Sweden
Clemens Dubslaff, Eindhoven University of Technology, Netherlands
Valentina Franzoni, Università degli Studi di Perugia, Italy
Pitoyo Hartono, Chukyo University, Japan
Thomas Hellström, Umeå University, Department of Computing Science, Sweden
Dmitry Kangin, Lancaster University, United Kingdom
Enea Parimbelli, University of Pavia, Italy
Karinne Ramírez-Amaro, Chalmers University of Technology, Sweden

(list not yet complete)

PAPER SUBMISSION

Prospective authors are invited to submit papers in any of the topics listed above.
Instructions for preparing the manuscript (in Word and Latex formats) are available at: Paper Templates
Please also check the Guidelines.
Papers must be submitted electronically via the web-based submission system using the appropriated button on this page.

PUBLICATIONS

After thorough reviewing by the special session program committee, all accepted papers will be published in a special section of the conference proceedings book - under an ISBN reference and on digital support - and submitted for indexation by SCOPUS, Google Scholar, DBLP, Semantic Scholar, EI and Web of Science / Conference Proceedings Citation Index.
SCITEPRESS is a member of CrossRef (http://www.crossref.org/) and every paper is given a DOI (Digital Object Identifier).
All papers presented at the conference venue will be available at the SCITEPRESS Digital Library

SECRETARIAT CONTACTS

ICAART Special Sessions - IAI 2025
e-mail: icaart.secretariat@insticc.org
footer