Simone Cuconato

Logic and theoretical computer science are deeply interconnected, with logic forming a foundational pillar in the emergence of computer science. This connection has grown stronger over time, driven by advancements in symbolic systems within artificial intelligence, formal verification methods, and automated reasoning techniques. However, within the vast landscape of computer science, data science remains an area where the link to logic is relatively underdeveloped. In this paper, I apply algebraic logic to address an urgent challenge for data scientists: metadata veracity. Inspired byWillard Van Orman Quine’s well-known slogan, “No entity without identity”, I propose, “No data without metadata”, underscoring the importance of descriptive metadata in scientific articles. The central idea of my framework is that the geometric zones of a scientific article – with its atomic metadata –, correspond to an algebraic situation space. Consequently, I develop a situation-based semantics grounded on the idea that situations are portions of a possible world, a possible world is a PDF document and that sentences point to situations and describe what is going on in them. Applying mathematical logic tools to data mining enables a rigorous framework for defining metadata veracity, offering a structured approach to assess the accuracy and reliability of extracted information.

There has been much discussion of the indispensability argument for the existence of mathematical objects. However, there has been little discussion of the relationship between Meinongianism and the indispensability argument. According to strong Meinongianism: i) some objects do not exist; ii) we can refer to and quantify over nonexistent objects to make true statements about them; and iii) nonexistent objects are the truthmakers of sentences which contain reference to or/and quantification over nonexistent objects. In this paper, against the mainstream Quinean meta-ontology, I argue that there are nonexistent mathematical objects as truthmakers and that nonexistent mathematical objects make a difference for the way the concrete world is.

In an increasingly data-driven world, the question of data – or metadata – veracity is now a central issue not only in the world of information but also in the legal one. Data veracity describes a closeness to truth on a higher level than a measure such as accuracy does. High veracity data is data that can be relied upon when making decisions, thus reducing the risk of basing choices on untrue information. The article uses epistemic logic to model structured metadata automatically extracted from legal papers, and the tools of metaontology to propose a definition of veracity as truthmaker.

The field of epistemic logic developed into an interdisciplinary area focused on explicating epistemic issues in, for example, artificial intelligence, computer security, game theory, economics, multiagent systems and the social sciences. Inspired, in part, by issues in these different ‘application’ areas, in this paper I propose an epistemic logic T for metadata extracted from scientific papers on COVID-19. More in details, I introduce a structure S to syntactically and semantically modelling metadata extracted with systems for extracting structured metadata from scientific articles in a born-digital form. These systems will be considered, in the logical model created, as ‘Metadata extraction agents’. In this case MEA taken into consideration are CERMINE and TeamBeam. In an increasingly data-driven world, modelling data or metadata means to help systematise existing information and support the research community in building solutions to the COVID-19 pandemic.

In this article we develop a logical model for automatic extraction of structured metadata. We introduce a new predicate???? – reads ‘extract’ – and a structure???? to syntactically and semantically define metadata extracted with any automatic metadata extraction system. These systems will be considered, in the logical model created, as knowledge extraction agents. In this case KEA taken into consideration is CERMINE, a comprehensive open-source system for extracting structured metadata from scientific articles in a born-digital form.