Date	October 1, 2023
Location	The workshop will be held in-person at the ECAI 2023 conference in Kraków, Poland.
Room	A-1-13 in the Faculty of Physics, Astronomy and Applied Computer Science of the Jagiellonian University

Invited Speakers [Abstracts]


Concha Bielza Technical University of Madrid	Pedro Larrañaga Technical University of Madrid	Issam El Naqa Moffitt Cancer Center	Christin Seifert University of Marburg

Schedule

This is the workshop’s schedule in Central European Time (CET).

Time (CET)	Event
09:00 - 09:10	Opening remarks
	[Session 1]
09:10 - 10:00	Invited talk by Christin Seifert: Can we trust XAI? Current status and challenges of evaluating XAI methods. [slides]
10:00 - 10:15	Paper: SHAP values from a Practical Perspective of a Social Scientist. Łukasz Borowiecki.
10:15 - 10:30	Paper: Evaluation of Local Model-agnostic Explanations: Taxonomy and Limitations. Amir Hossein Akhavan Rahnama.
10:30 - 11:00	Coffee break
	[Session 2]
11:00 - 11:50	Invited Talk by Concha Bielza & Pedro Larrañaga: Explanation Capabilities of Bayesian Networks in Dynamic Industrial Domains. [slides]
11:50 - 12:05	Paper: Explainable Anomaly Detection in Industrial Streams. Jakub Jakubowski, Przemysław Stanisz, Szymon Bobek, Grzegorz Nalepa.
12:05 - 12:20	Paper: Clash of the Explainers: Argumentation for Context-Appropriate Explanations. Leila Methnani, Andreas Theodorou, Virginia Dignum.
12:30 - 13:30	Lunch break
	[Session 3]
13:30 - 14:20	Invited Talk by Issam El Naqa: Towards Trustworthy AI for Clinical Oncology. [slides]
14:20 - 14:35	Paper: From Data to Insights: Fusing Explainable AI and Epidemiological Thinking to Enhance Primary Care Practices. Awais Ashfaq, Slawomir Nowaczyk.
14:35 - 14:50	Paper: EcoShap: Save Computations by Only Calculating Shapley Values for Relevant Features. Parisa Jamshidi, Slawomir Nowaczyk, Mahmoud Rahat.
15:00 - 15:30	Coffee break
	[Session 4]
15:30 - 15:45	Paper: Towards Explainable Deep Domain Adaptation. Szymon Bobek, Zahra Taghiyarrenani, Sepideh Pashami, Slawomir Nowaczyk, Grzegorz Nalepa.
15:45 - 16:00	Paper: Interpretability benchmark for spatial alignment of prototypical parts. Mikołaj Sacha, Bartosz Jura, Dawid Rymarczyk, Łukasz Struski, Jacek Tabor, Bartosz Zieliński.
16:00 - 16:15	Paper: Evaluation of Human-Understandability of Global Model Explanations using Decision Tree. Adarsa Sivaprasad, Ehud Reiter.

Call for Papers

Welcome to the Joint workshops on XAI methods, challenges and applications (XAI^3), where we aim to discuss opportunities for the new generation of explainable AI (XAI) methods that are reliable, robust, and trustworthy. Explainability of AI models and systems is crucial for humans to trust and use intelligent systems, yet their utility in high-risk applications such as healthcare and industry has been severely limited. Our workshop will have three tracks: medical, industry, and future challenges, where we will explore the challenges and opportunities in creating useful XAI methods for medical applications, integrating explainability in highly automated industrial processes, and evaluating current and future XAI methods. We welcome contributions from researchers, academia, and industries primarily from a technical and application point of view, but also from an ethical and sociological perspective. Join us in discussing the latest developments in XAI and their practical applications at the 26th European Conference on Artificial Intelligence (ECAI 2023) in Kraków, Poland.

Tracks and topics

Towards Explainable AI 2.0 (XAI2.0) Chair: Przemysław Biecek	Explainable AI for Medical Applications (XAIM) Chair: Neo Christopher Chung	XAI for Industry 4.0 & 5.0 (XAI4I) Chair: Sławomir Nowaczyk
Emerging challenges in explainable AI towards XAI 2.0 Evaluation and limitations of current XAI methods Trade-off between model-agnostic and model-specific explainability Adversarial attacks and defenses in XAI Privacy, leakage of sensitive information, fairness and bias Human-centered XAI through visualization, active learning, model improvement and debugging XAI beyond classification and regression, e.g. in unsupervised learning, image segmentation, survival analysis	Theory and application of XAI for medical imaging and other medical applications Uncertainty estimation of AI models using medical data Multimodal learning, e.g., PET/CT, healthcare records, genomics, and other heterogeneous datasets Clinical cases, evaluation, and software of XAI for medicine Fairness, bias, and transparency in medical AI models Human-computer interaction (HCI) and human in the loop (HITL) approaches in medicine Inherently interpretable models in supervised, unsupervised and semi-supervised learning for biology and medicine	Ethical considerations in industrial deployment of AI AI transparency and accountability in smart factories Explainable systems fusing various sources of industrial information XAI in performance and efficiency of industrial systems Prediction of maintenance, product, and process quality Data and information fusion in the industrial XAI context Application in manufacturing systems, production processes, energy, power, and transport systems

Dates

Paper submission deadline	~~July 18, 2023~~
Decision notification	~~August 15, 2023~~
Camera-ready due	~~September 11, 2023~~

All times Anywhere on Earth (AoE), UTC-12

Submission instructions

Submissions should be anonymised and follow the LNCS template available as downloadable LaTeX and Word, as well as online Overleaf template, at https://www.springer.com/gp/computer-science/lncs/conference-proceedings-guidelines

We welcome two types of papers:

Non-proceedings track: We welcome short papers of 4-8 pages. Submissions accepted to this track will be presented at the XAI^3 workshop but not published in Springer proceedings.
Proceedings track: We welcome full research papers of 10-15 pages. Full papers presented at the XAI^3 workshop will be published by Springer Nature in joint ECAI workshop proceedings volume, as part of the CCIS book series. It will be possible to make individual papers Open Access, at the discretion and cost of the authors, by following the Springer procedure described here.

The page limit does not include references and supplementary material.

Submissions can be done at https://cmt3.research.microsoft.com/XAI3ECAI2023

Organisers


Hubert Baniecki University of Warsaw	Przemysław Biecek Warsaw University of Technology	Albert Bifet Telecom-ParisTech	Szymon Bobek Jagiellonian University	Lennart Brocki University of Warsaw	Giuseppe Casalicchio Ludwig Maximilian University of Munich


Neo Christopher Chung University of Warsaw	Joao Gama University of Porto	Mathieu Hatt LaTIM	Grzegorz J. Nalepa Jagiellonian University	Sławomir Nowaczyk Halmstad University	Panagiotis Papadimitroulas BIOEMTECH


Sepideh Pashami Halmstad University	Rita P. Ribeiro University of Porto	Dawid Rymarczyk Jagiellonian University	Jacek Tabor Jagiellonian University	Bruno Veloso University of Porto	Bartosz Zieliński Jagiellonian University