What can FCA do for Artificial Intelligence?

Based on the success of the preceding editions, the three organizers of FCA4AI Workshop Series have decided to continue the adventure and to propose a 12th edition of the FCA4AI workshop co-located with ECAI 2024 in Santiago de Compostela (Spain, October 2024). The preceding editions of the workshop (from ECAI 2012 until IJCAI 2023) have shown that there are many AI researchers interested in FCA and, in parallel, many researchers in FCA are actually working in AI. The scientific community interested in FCA and related research work is mostly based in Europe, and thus this is a good opportunity to propose a new edition of the FCA4AI workshop at next ECAI 2024 in Spain.

Formal Concept Analysis (FCA) is a mathematically well-founded theory aimed at data analysis and classification. FCA allows on to build a concept lattice and a system of dependencies (implications and association rules) which can be used for many AI needs, e.g., knowledge processing, knowledge discovery, knowledge representation and reasoning, ontology engineering, as well as information retrieval, recommendation, social network analysis and text processing. Thus there are many ``natural links'' between FCA and AI.

Recent years have been witnessing increased scientific activity around FCA, in particular a strand of work emerged that is aimed at extending the possibilities of FCA w.r.t. knowledge processing, such as work on pattern structures and relational concept analysis. These extensions are aimed at allowing FCA to deal with more complex than just binary data for solving complex problems in knowledge and data processing. While the capabilities of FCA are extended, new possibilities are arising in the framework of FCA.

As usual, the FCA4AI workshop is dedicated to the discussion of such issues, and in particular:

• How can FCA supports AI activities in knowledge discovery, knowledge representation and reasoning, machine learning, natural language processing...
• By contrast, how the current developments in AI can be integrated within FCA to help AI researchers solve complex problems in their domain,
• Which role can be played by FCA in the new trends of AI, especially in ML, XAI, fairness of algorithms, and ``hybrid system'' combining symbolic and subsymbolic approaches.

Topics of interest

Topics of interest include but are not limited to:

• Concept lattices and related structures: pattern structures, relational structures, distributive lattices.
• Knowledge discovery and data mining: pattern mining, association rules, attribute implications, subgroup discovery, exceptional model mining, data dependencies, attribute exploration, stability, projections, interestingness measures, MDL principle, mining of complex data, triadic and polyadic analysis.
• Knowledge and data engineering: knowledge representation, reasoning, ontology engineering, knowledge graphs, mining the web of data, text mining, data quality checking.
•  Analyzing the potential of FCA in supporting hybrid systems: how to combine FCA and data mining algorithms such as deep learning for building hybrid knowledge discovery systems, producing explanations, and assessing system fairness. 
• Analyzing the potential of FCA in AI tasks such as classification, clustering, biclustering, information retrieval, navigation, recommendation, text processing, visualization, pattern recognition, analysis of social networks.
• Practical applications in agronomy, astronomy, biology, chemistry, finance, manufacturing, medicine...

The workshop will include time for audience discussion aimed at better understanding of issues, challenges, and ideas being presented.