PhD position on Data Diversity for Fair and Robust NLP (DataDivers project)

Are you passionate about fair and robust Natural Language Processing (NLP), data, and computational social science/sociolinguistics? Join our new ambitious DataDivers project funded by an ERC Starting grant and help us make NLP models more fair and robust.

Your job
The rise of Large Language Models (LLMs) and the availability of massive datasets have sparked a revolution in the field of NLP. However, numerous studies have pointed towards serious flaws: NLP models encode societal biases and show disparate performance across demographic groups. Thus, current models can and do cause real harm when deployed in society.

In the field of NLP, there is a growing recognition that data quality is key to better language models, yet we know surprisingly little about the link between data and model behaviour. In this project, we will develop methods to measure the diversity of NLP datasets, assess the impact of diversity on NLP models, and improve data collection and model training.

As a PhD candidate in our new DataDivers project, you will join the project team led by Dr Dong Nguyen. The team will consist of two PhD candidates and two Postdocs.

You will develop innovative methods to measure the diversity of NLP datasets. A major focus will be on measuring the dataset diversity from a sociolinguistic perspective, considering language variation – such as styles and dialects - and combining (socio)linguistic insights with neural language modelling. You will also draw from relevant disciplines, particularly the social sciences, that have developed measurement approaches for diversity. Furthermore, you will carry out experiments to assess the impact of data diversity on NLP models, with a focus on fairness and robustness, and investigate ways to leverage data diversity to improve NLP models!

This position offers you the opportunity to work on fundamental NLP research. As a PhD candidate, you will have the freedom to shape the project according to your own interests. Responsibilities include contributing to teaching activities, such as supervising Bachelor’s and Master’s theses or assisting in labs.

Requirements:

We are looking for an ambitious and collaborative PhD candidate, who meets several or all of the following criteria:

• You hold an MSc degree in Artificial Intelligence, Natural Language Processing, Machine Learning, Linguistics, Computational Social Science, or a related field, with demonstratable experience in NLP and Machine Learning.
• You have strong programming skills. Experience with High-Performance Computing, or a willingness to learn, is desirable.
• You have a multidisciplinary mindset and are proactive in exploring and integrating knowledge from various fields.
• You have excellent written and verbal communication skills in English.
• You have strong teamwork skills, as you will be collaborating closely with the larger DataDivers team.

Salary Benefits:

We offer:

• a position for four years;
• a gross monthly salary between €2,872 and €3,670 in the case of full-time employment (salary scale P under the Collective Labour Agreement for Dutch Universities (CAO NU));
• 8% holiday pay and 8.3% year-end bonus;
• a pension scheme, partially paid parental leave and flexible terms of employment based on the CAO NU.

In addition to the terms of employment laid down in the CAO NU, Utrecht University has a number of schemes and facilities of its own for employees. This includes schemes facilitating professional development, leave schemes and schemes for sports and cultural activities, as well as discounts on software and other IT products. We also offer access to additional employee benefits through our Terms of Employment Options Model. In this way, we encourage our employees to continue to invest in their growth. For more information, please visit Working at Utrecht University.

36 - 40 hours per week

Budapestlaan 6