In an era dominated by data-driven insights, designing and building trustworthy Computational Intelligence (CI) systems presents fundamental challenges, especially when leveraging data from decentralized sources. Privacy regulations and ethical concerns have raised the demand for CI systems that are both collaborative and reliable, all while upholding stringent standards of data privacy and transparency. Federated Learning (FL) has emerged as a groundbreaking solution to address these challenges. By allowing multiple parties—such as institutions, corporations, or devices—to collaboratively train models without directly sharing their data, FL provides a paradigm shift that enables decentralized learning while preserving data privacy. This approach minimizes the risk of exposing sensitive information and aligns well with evolving regulatory frameworks, such as the GDPR in Europe, which prioritizes data sovereignty.

However, to truly achieve trustworthiness in CI, ensuring privacy is only one piece of the puzzle. Explainability and transparency are equally critical, as they allow end-users and stakeholders to understand and trust the decisions made by these systems. The need for model interpretability becomes particularly complex in a Federated Learning setting, where data resides on diverse, decentralized nodes. Typically, explainability is achieved either by designing inherently interpretable models or through post-hoc interpretability techniques that elucidate model behavior after training. Yet, when these techniques are applied in the FL context, maintaining both privacy and interpretability poses significant technical and ethical challenges.

The aim of this special session is to provide a multidisciplinary and international forum to discuss emerging methodologies for the collaborative learning of CI systems that are both private and trustworthy. This session seeks to explore advanced FL techniques that incorporate robust privacy measures alongside mechanisms for explainability, with the overarching goal of fostering CI models that users and stakeholders can trust. Attendees from various fields—such as machine learning, ethics, law, and policy—are invited to contribute perspectives and solutions that address the unique challenges of building decentralized, privacy-preserving, and interpretable CI systems.

This Special Session is supported by the project “FAIR - Future Artificial Intelligence Research” - Spoke 1 - funded by the European Union under the NextGenerationEU programme.

The topics of interest include (but are not limited to):

• Trustworthy Collaborative Artificial Intelligence

• Collaborative/Federated Learning of Explainable Artificial Intelligence Models

• Collaborative/Federated Learning of Interpretable by-design models

• Collaborative/Federated Learning and post-hoc explainability techniques 

• Collaborative/Federated Learning of  Trustworthy Large Language Models

• Collaborative/Federated Supervised, Semi-Supervised and Unsupervised Learning for Fuzzy and Neuro-Fuzzy Models

• Collaborative/Federated Supervised, Semi-Supervised and Unsupervised Learning for Neural Networks and Deep Learning Models 

• Neural Networks and Fuzzy sets theory for Collaborative/Federated Learning

• Privacy Preserving Machine Learning

• Threats, attacks and defenses to Federated Learning

• Applications of Trustworthy Collaborative Artificial Intelligence

Submitted papers will be peer-reviewed with the same criteria of other IJCNN tracks.

The papers accepted for the special session will be included in the IJCNN 2025 proceedings and will be published by the IEEE Xplore Digital Library.

Organizing Committee

• Pietro Ducange, University of Pisa, Italy

• Michela Fazzolari, Istituto di Informatica e Telematica, CNR, Italy

• Francesco Marcelloni, University of Pisa, Italy

• Alessandro Renda, University of Pisa, Italy

Publicity and Communication Chair:

• José Luis Corcuera Bárcena, University of Pisa, Italy

Corresponding Chair's E-mail:

• pietro.ducange@unipi.it

• CLOTHES 2024@WCCI 2024