INSAIT: INterpretable Systems for Artificial Intelligence Transparency

ICIAP 2025 Workshop

About

Artificial Intelligence (AI) innovations have led to increasingly complex models, such as foundational and generative multimodal models. They achieve state-of-the-art performance across various domains, however, their black-box nature raises concerns about their trust, accountability, and ethical deployment. The INSAIT Workshop (INterpretable Systems for Artificial Intelligence Transparency) aims to bring together researchers and practitioners to explore interpretability in AI, focusing on methods that enhance transparency, explainability, and human-centric trust. Understanding why AI systems work so well and why they fail is important for their integration into real-world applications used by millions of people.

We are now accepting self-nominations for reviewers, please complete this form.

Call for papers

This workshop invites contributions from academia and industry on theoretical and practical topics including, but not limited to:

  • Explanation methods: post-hoc interpretability techniques (e.g., SHAP, LIME, attention-based explanations, mechanistic interpretability, etc.).
  • Interpretability by design: design of inherently interpretable architectures (e.g., prototype-based, capsule networks, concept bottleneck models).
  • Concept learning and unlearning techniques: for example, how removing unwanted concepts impacts a model.
  • Human-AI interaction: how interpretability improves user trust and decision-making.
  • Benchmarking interpretability: standardized evaluation metrics for explainability.
  • Ethical & societal implications: ensuring fairness, accountability, and transparency through interpretability.
  • Case studies: applications of interpretability in healthcare, finance, legal systems, and autonomous systems.

Paper submissions

All submissions will go through a double-blind review process. Papers will be selected based on relevance, significance, novelty of results, technical merit, and clarity of presentation. Only previously unpublished works will be considered.

There are two tracks for submission to INSAIT:

Proceedings track: Submitted papers to the proceedings track must not exceed 12 pages (including references). Accepted papers will be published in a dedicated volume of Springer Lecture Notes in Computer Science (LNCS).

Extended abstract track: In this track, we invite papers up to 6 pages (including references) for presenting high-impact, innovative ideas or work-in-progress that may not be ready for full publication.

Submissions to both tracks will be featured during the workshop’s poster session, and a subset of all submissions will be selected for spotlight talks. Papers must be in English.

Each accepted paper must be covered by at least one author registration (either a Full registration or a Workshop/Tutorial registration if you plan to attend the workshops/tutorials only).

We suggest workshop papers are prepared and submitted using the official ICIAP template (there is also this Overleaf template).

Please submit your papers in PDF format through OpenReview.

Important Dates

Submission deadline - June 7, 2025
Author notification - July 01, 2025
Camera ready deadline - July 10, 2025

Note: all deadlines are in Anywhere on Earth (AoE).

Workshop Event

When: September 15 or 16 (to be announced), 2025
Where: Sapienza University of Rome, Piazzale Aldo Moro 5, 00185 Rome

Registration

For detailed instructions and information on registration fees, please visit the ICIAP registration page.

Schedule

To be announced.

Speakers

Organizers

Program Committee

Marco Grangetto, University of Turin
Enrico Cassano, University of Turin
Muhammad Rashid, University of Turin
Marco Nurisso, Politecnico di Torino
Georgios Leontidis, University of Aberdeen
Aiden Durrant, University of Aberdeen
Sonia Laguna Cillero, ETH Zurich
Eliana Pastor, Politecnico di Torino
Salvatore Greco, Politecnico di Torino
Lia Morra, Politecnico di Torino
Giulia Vilone, Analog Devices
.. (to be updated)

Contact

Organized by

UNITO Logo POLITO Logo UNIPI Logo CentAI Logo