Seminari di Lukasz Gajewski e Francesco Monzillo

Dr. Bot: A chatbot for medical self-diagnosis di Lukasz Gajewski

Abstract: In recent years, the use of ‘Dr Google’ as a medical self-diagnosis tool has exponentially surged in popularity. However, while “Dr. Google” can be a valuable resource for medical self-diagnosis, it has its limitations.
This is why having a reliable medical self-diagnosis tool is important in today’s digital age. It allows individuals to take an active role in their own healthcare and have immediate access to a healthcare professional is not available.
Chatbots are particularly useful for this purpose as they simplify human-machine interaction and can understand and respond to user questions and requests intuitively. The thesis aims to develop an open source chatbot that allows users to describe their condition in natural language. The chatbot will then extrapolate the user's assumed symptoms and continue the diagnosis process based on the partial results, guiding the user towards the most appropriate diagnosis.
The chatbot will utilize knowledge graphs as a data source, adhering to the rules of the semantic web, because they are based on well-archived information from reliable sources.
The thesis will conduct a comprehensive analysis of the available knowledge graphs within the medical field using the Linked Open Data (LOD) Cloud search tool to find published knowledge graphs and analyzing literature indexed on Scopus to evaluate the latest research in the field of knowledge graphs.
This research highlighted the limited availability of medical resources in the form of standardized Knowledge Graphs, in fact the research produced only 11 knowledge graphs that could be used for my project.
Upon evaluating the chatbot tool, a comprehensive performance analysis will be conducted, comparing it directly to ChatGPT, a rapidly growing platform that has the potential to dominate the market as a new ‘Dr. Google’. This comparison will provide insights into the current state of Dr. Bot and how it should position itself among the emerging alternatives.

Valutazione di MPPCO (Medieval Philosophy works Publishing And Commenting Ontology) un'Ontologia basata su Screenia di Francesco Monzillo

Abstract: Un'analisi di MPPCO a partire dai concetti, requisiti sui quali è fondata e specifica dei dettagli di progettazione sottostanti fino alla descrizione degli aspetti riguardanti la valutazione della soluzione proposta. Il seminario procederà con la presentazione delle metodologie di valutazione, i suoi partecipanti, i risultati della valutazione e una discussione finale riguardante le limitazioni della soluzione e le sue possibili direzioni future.