This assignment is a close collaboration between a national AI lab and a Natural Language Processing group within a research institution. The goal is to explore how small models can extract useful information on specific crime themes, such as drug trafficking, primarily from publicly available news sources. Relevant tasks include identifying relevant entities, inferring timelines, and relations.
A promising direction is to train or fine-tune small-scale models via distillation from larger models, including through the use of synthetic data. This in turn raises interesting questions about the robustness of the resulting models, their ability to generalise, and their susceptibility to known issues related to synthetic data training.
What you will do.
- Design and evaluate methods for multimodal entity, timeline, and relation extraction from multilingual news sources.
- Fine-tune and distil small-scale language models from larger models, including via synthetic data generation.
- Investigate the robustness, generalisability, and limitations (e.g., model collapse) of the resulting models.
- Apply and validate the approach on real-world use cases, such as drug trafficking, together with a national police organization.
- Publish and present your findings at relevant conferences and in scientific journals.
Profile
- A PhD (obtained, or nearly completed) in Natural Language Processing or a closely related area.
- Solid knowledge of machine learning, especially deep learning.
- Experience in model development and/or fine-tuning.
- A practical mindset – interested in seeing your models used, not just published.
- Fluency in spoken and written English (C1 level).
- Good knowledge of Dutch is an asset, though not required, to facilitate collaboration with the national police organization.
How to apply
View the full assignment text and application details once your tailored application is ready.
Order a tailored application to view the full assignment and application details.
More context, less searching.
You get enough context to judge whether this job is relevant. The full brief, client details and next steps stay available inside the app.