Postdoctoral researcher in Artificial Intelligence: Learning from Sparse Examples in Natural Language: 1.0 FTE

Project and background

Tilburg University | Tilburg School of Humanities and Digital Sciences, Department of Cognitive Science and Artificial Intelligence is looking for a Postdoc within the HAICu research project. The HAICu project is a large-scale Dutch research project by universities and cultural-heritage institutions into new forms of Artificial Intelligence-based access to multimodal Cultural-Heritage data, both contemporary and historical. Within HAICu, AI researchers, Digital Humanities researchers and a wide range of public and private partners will co-develop scientific solutions to unlock the true societal potential of the current heterogeneous digital heritage collections. It will provide easier, richer and more reliable data access to citizens, journalists, civic organizations, and various other stakeholders.

To this end, the HAICu project will  research new forms of Artificial Intelligence-driven access to multi-modal data stored in Dutch cultural heritage institutions. Within HAICu, Artificial Intelligence researchers and Digital Humanities researchers will co-develop a scientific solution to unfold the true potential of the current heterogeneous digital heritage collections and provide easier and reliable data access to citizens, journalists, civic organizations, and other societal users. 

HAICu is funded by the NWO National Science Agenda (NWA) and has a budget of about EUR 10 million. HAICu started in February 2024 and will last 6 years (until Jan 2030).

The postdoc  will participate in a Work Package (3) concerning Learning from Sparse Examples,  in collaboration with RUG, Naturalis, Fontys Hogeschool, National Library (Nationale Bibliotheek, KB), Sound and Vision (Beeld & Geluid) and  develop advanced machine-learning methods that can deal with sparse and possibly multimodal data with Geometric Deep Learning Approaches on sparse textual data from Dutch National library.  The sparsity may be in the limited number of instances available or in the limited number of labels available. 

The postdoc will be primarily based in Tilburg but work closely with our RUG partners in Groningen and with the National Library in The Hague. Travelling funds for both locations are included in the project budget.

In addition to the publication of scientific articles, the researcher is expected to make the developed algorithms available to academic and institutional project partners, under open access conditions.

As a postdoc researcher:

You will be positioned within a WP where the task is to  develop advanced machine-learning methods that can deal with sparse and possibly multimodal data. The sparsity may be in the limited number of instances available or in the limited number of labels available. Dealing with sparse data will be achieved by means of (i) transfer learning, (ii) structured approaches such as Graph Convolutional Networks, and (iii) unsupervised, self-supervised and semi-supervised methods.

As there is expected coherence and interaction between the work packages of this project, you will be in contact with representatives of other work packages.

The algorithmic tools will be developed according to the “expert-in-the-loop” principle. Each version of the algorithms and tools will be presented to experts and evaluated by them. In this case, “experts” refers to both domain specialists and citizen specialists.

More specifically your research will be linked to the following tasks:

• selecting relevant graph-based deep learning and knowledge-graph embedding methods for semantic network induction with application domain the monitoring of (news) data stream
• performing large-scale experiments to assess to what extent the provenance of news can be inferred. Statistical evaluation of results and identifying bottlenecks.

Requirements:

Requirements of the Postdoc is to: 

• be a motivated researcher, with a PhD in machine learning, AI, computer science,  data science, computational linguistics, information science or a comparable domain.
• have experience with geometric deep learning/graph representation learning, explainable AI methods
• have ability to conduct high quality academic research, reflected in demonstratable outputs.
• Be  a team player proven ability to manage complex projects .
• Have the ability to collaborate effectively in diverse teams; someone who is willing to and able to do  communication/ administration of connections across stakeholders is preferred.
• have excellent (written and verbal) proficiency in English, good communication and leadership skills.
• Affinity with the Dutch language, cultural context and history are pluses. 

Salary Benefits:

Conditions of employment

Tilburg University offers excellent benefits in a pleasant working environment: 

• A position based on 0.8 - 1.0 FTE (32-40 hours per week). 
• A salary of minimum € 4,537 and maximum €6,209 gross per month for full-time employment, based on UFO profile Researcher 3 and salary scale 11. Until obtaining your PhD, salary scale 10 will be applied. Tilburg University uses a neutral remuneration system based on relevant work experience. 
• This is a vacancy  for a temporary position in accordance with Article 2.3 paragraph 5 sub b CLA DU. You will be given a temporary contract for the duration of 24 months.
• Vacation pay of 8% and a year-end bonus of 8.3%. 
• Over 8 weeks of vacation leave. 
• The opportunity to work partly on campus and partly from home with a home office allowance of €2 per day.
• Reimbursement for sustainable commuting: walking, cycling, and public transport. 
• A monthly internet allowance of €25.
• An options model in which you exchange benefits for things such as additional leave, more pension, a bicycle or personal training at our Sports Center. 
• A moving allowance (subject to conditions). 
• Candidates from outside the Netherlands may qualify for a tax-free allowance. During the first 20 months, 30% of the salary will be reimbursed untaxed. In the next 20 months, this percentage drops to 20% and during the last 20 months, the untaxed reimbursement is 10% of the salary. The university will apply for such an allowance on their behalf.
• A pension with ABP; the largest Dutch pension fund.
• Training in personal development, career development, leadership, education, and research. Or a language course at our Language Center. 
• A work culture in which we embrace differences, everyone is welcome and given equal opportunities. 
• A vibrant campus in green surroundings that is easily accessible by public transport. 

For more information, see our website and the CLA Dutch Universities.   

32 - 40 hours per week

Warandelaan 2