Tangles: a structural approach to artificial intelligence in the empirical sciences (Part I)
This foundational method addresses the problem of fuzzy clustering and type discovery in empirical sciences, potentially impacting AI applications across domains.
The paper introduces tangles as a structural approach to identify groups of qualities that co-occur, enabling discovery and structuring of types in various domains, and offers a precise paradigm for fuzzy clustering without requiring object assignments.
Traditional clustering identifies groups of objects that share certain qualities. Tangles do the converse: they identify groups of qualities that often occur together. They can thereby discover, relate, and structure types: of behaviour, political views, texts, or viruses. If desired, tangles can also be used as a new method for traditional clustering. They offer a precise, quantitative paradigm suited particularly to fuzzy clusters, since they do not require any assignment of objects to the clusters which these collectively form. This is the first of four parts of a book with the above title. The book explores applications outside mathematics of the notion and theory of tangles generalised from the graph tangles know from graph minor theory.