Special Issue

Explainable and Evolving AI Systems (EEAIS)

A Special Issue of

Evolving Systems – An Interdisciplinary Journal for Advanced Science and Technology Science and Technology

Guest Editors

Giovanna Castellano, Edwin Lughofer, George A. Papadopoulos.

Overview

Recent rapid advances in Artificial Intelligence (AI) are tied largely to evolving systems that provide remarkable benefits in terms of accurate predictions and reliable decision making. However, beyond high accuracy, it is essential for new AI systems to be understandable by users, who should be able to explain the behavior of AI systems and effectively manage them. Hence, engineers and scientists developing AI systems are focusing more and more their attention on learning explainable models from data that will enable them to gain confidence with the decisions of AI systems and, as a consequence, trust them. More generally, the importance of learning explainable models from data is of help to (i) improve the interaction between users and AI systems in order to tackle complex problems, (ii) easily integrate artificial and human knowledge, and (iii) allow users to validate an AI system with respect to criteria of performance, ethics, safety, causality, etc., thus leading to the ultimate possibility of trusting AI systems for mission-critical applications. Evolving and adaptive models play a fundamental role in the development of AI systems which have the distinction of leveraging explainable knowledge that emphasizes transparency, interpretability and scalability. Evolving models comprise an array of online modeling approaches capable of extracting knowledge from online data streams generated by nonstationary processes. They embody online learning methods and incremental algorithms that evolve or gradually change the extracted knowledge to guarantee life-long learning and self-organization of the granular structure of the model. 

Topics

The special issue is intended to focus on the above aspects and will solicit papers that cover original research, overviews and applications of evolving methods in the realm of explainable AI systems.

Areas of interest include, but are not limited to:

  • Evolving explainable models for Data Streams and/or Big Data 
  • Evolving models for Explainable Artificial Intelligence 
  • Evolving solutions for real-time explainable models 
  • Neural/Neuro-fuzzy networks to derive explainable and evolving models
  • Granular Computing and Fuzzy systems for Explainable and Evolving models
  • Adaptive personal explainable systems 
  • Explainable models for behavior analysis 
  • Explainable and evolving models for image processing
  • Explainable and evolving methods for user profiling, opinion mining and sentiment analysis 
  • Incremental Learning of explainable models for texts and document mining 
  • Explainable and evolving methods for Human-computer interaction 
  • Explainable and evolving systems for e-health 
  • Explainable and evolving systems for smart cities 
  • Industrial and real-world applications of Explainable Evolving systems

Submission and Review Process

Authors of accepted papers at the 2022 IEEE International Conference on Evolving and Adaptive Intelligent Systems (IEEE EAIS 2022) are invited to submit an extended version of their paper.

In addition, any other high-quality submission that fits the topics of this special issue is welcome. All invited papers will be subjected to the same rigorous review process as the regular submissions to this special issue. Submitted articles must not have been previously published or currently submitted for publication elsewhere. For work that has been published previously in a workshop or conference, it is required that submissions to the special issue report substantial advancements in research and have at least 40% of new content.

Papers will be screened by the guest editors and those deemed suitable will be sent to at least two reviewers. Manuscripts must apply the general author guidelines of the Journal, which are available at (https://www.springer.com/journal/12530/submission-guidelines) and must be submitted through the journal’s online submission portal (https://www.editorialmanager.com/evos/default.aspx). 

Tentative Schedule

Submissions open: August 1, 2022 

Submissions deadline: December 15, 2022 

Notification: February 15, 2023 

Revision submission: March 15, 2023

Final notification: April 15, 2023 

Publication: 4th quarter of 2023