Machine Learning for Information Extraction from Ancient Texts

(ref. BAP-2022-522)

Laatst aangepast : 11/07/2022

Within the context of the ID-N project NIKAW, funded by the Special Research Fund (BOF) of KU Leuven, we are inviting applications for a PhD position on the subject of machine learning for information extraction from ancient texts. The set of thoughts, ideas, and beliefs that are shared, discussed, and enriched in a society is a complex and dynamic reality, constantly fed by and giving way to historical events. Nowadays, the easy access to social network and telecommunication data, online publications, and, for instance, book sales data, has increased our possibilities of analyzing this flow of information incredibly. In this project, however, we want to investigate how we can map the circulation of ideas and identify pivotal individuals in past societies, namely the Graeco-Roman world, whose fragmented history, encompassing many centuries, can only be studied through partial remainders. To this end, the NIKAW project aims to exploit textual information from the ancient world to reconstruct the transmission of knowledge across multilingual, geographically and chronologically extended communities.

Project

The PhD candidate will focus on the extraction of relevant information from Ancient Greek and Latin texts, which will subsequently be used for further analyses within the project. Part of the tasks will be carried out in collaboration with another PhD student and one Postdoc. Research will be carried out along three tracks:

• Firstly, the candidate will develop a state of the art natural language processing pipeline for named entity extraction and named entity linking, specifically tailored towards classical texts. 
• Secondly, the candidate will develop a bilingual language model (Ancient Greek-Latin), based on neural NLP methods, that will facilitate named entity extraction and linking, and the representation of the context of the mentions in the corpus used.
• Thirdly, the candidate will improve the natural language processing pipeline by making use of additional information developed by other project members. This includes knowledge on potential linguistic and personal relations gained from social network analysis and learned network embeddings.

The candidate is expected to start in November or a soon as possible afterwards and, depending on their profile,

 will be supervised by one or more project (co)PIs.

Profile

• Although not restricted to these, you preferably hold a Master in Linguistics, Classical Studies, Computer Science, or equivalent education
• You have experience with machine learning, especially with neural networks (deep learning) for natural language processing
• You have an excellent proficiency in English and good oral and written communication skills
• Knowledge of a classical language (Latin and/or Ancient Greek)  is a plus
• You are a team player and enjoy collaborative and interdisciplinary projects

Offer

• We offer a full time PhD position for 1 year, extendable to 4 years after initial positive evaluation
• You will be able to conduct scientific research within a high-level research environment, leading to a doctoral degree
• You will work in a larger project, in cooperation with an interdisciplinary team of researchers (Margherita Fantoli, Mark Depauw, Alek Keersmaekers, Bart Thijs, Tim Van de Cruys, Toon Van Hal), and two KU Leuven Research Institutes (Leuven.AI and Lectio)
  
• You will have the opportunity to participate in international conferences, and benefit from academic training and workshops

Interested?

To apply, please upload a motivation letter, a CV and the contact details of two referees with your application. Selected candidates will be invited to an interview.

For more information please contact Prof. dr. Margherita Fantoli, tel.: +32 16 19 38 52, mail: [email protected] or Dr. Jens Bürger, tel.: +32 16 37 91 93, mail: [email protected].

KU Leuven seeks to foster an environment where all talents can flourish, regardless of gender, age, cultural background, nationality or impairments. If you have any questions relating to accessibility or support, please contact us at [email protected].