(Maddalena Zunino, Valentina Mignosa, Andrea Brunello, Marco Fucecchi, Nicola Saccomanno, Giuseppe Serra, Alessandro Tremamunno, Chiara Battistella)

The rich tapestry of Greek and Latin literature provides a fertile ground for the exploration of intertextuality, wherein texts echo, reference, or directly cite previous works, creating a dense web of linguistic and thematic connections. This project seeks to delve into the intricate realm of intertextuality within classic Greek and Latin texts, employing computer science techniques such as Natural Language Processing (NLP), text analysis, and machine learning. The goal is to unravel the complex network of thematic and semantic correlations, references, and echoes among various texts, thus providing a deeper understanding of the dialogues between different authors and works. Employing topic extraction, we aim to discern recurring themes and subjects that permeate diverse texts, offering insights into the pervading cultural and societal discourses of the time. Furthermore, by leveraging name entity recognition, we seek to map out the myriad of characters, places, and events that intertwine throughout Greek and Latin literature. The challenging linguistic aspects, such as metaphors and idiomatic expressions, can be addressed through deep learning and embeddings, enabling the model to discern semantic similarities beyond mere syntactic analogies. This exploration not only seeks to illuminate the multifaceted dialogues occurring within Latin and Greek literature but also to push the boundaries of what is achievable with computational linguistics and artificial intelligence in the realm of classical studies.

Tools and Technologies:

• NLP libraries: for instance, NLTK, SpaCy and Gensim libraries for Python
• Text analysis techniques: ranging from classical ones, such as LDA/LSA, to the usage of modern deep learning embeddings (PyTorch library for Python, pre-trained models such as Latin BERT and Ancient Greek BERT)
• Data Visualization:  Matplotlib and Seaborn libraries for Python, NetworkX
• Database Management: relational (Postgres) and non-relational solutions (graph databases: Neo4J, document-oriented stores: MongoDB)

Possibili coinvolgimenti nel progetto:

• Bachelor’s degree project: implement basic text pre-processing and analysis and NLP techniques to explore thematic parallels across selected Latin texts using classical tools, such as LDA and LSA
• Master’s degree project: dive deeper into specific techniques or aspects of intertextuality, possibly incorporating more advanced machine learning or deep learning models (e.g., Latin BERT embeddings)
• Internship: engage in a focused aspect of the project, such as data cleaning, basic analysis, or data management through a database, to gain hands-on experience with NLP and text analysis in a real-world project setting, contributing to a particular phase or aspect of the project
• PhD research project: engage in a thorough, in-depth exploration of intertextuality in Latin literature using and developing new advanced computational methods and database systems solutions, with the idea of creating a full-fledged framework to handle intertextuality in the classic Greek and Latin literature