Barbara Plank

Project title

MultiVaLUe: Multilingual Variety-aware Language Understanding Technology

What is your project about?

The goal of MultiVaLUe is to develop algorithms for information extraction enabling small languages to get a significant step closer to accuracy levels of well-supported languages like English. To do so, the project will devise extraction algorithms that work under limited (or lack of) training data.

How did you become interested in your particular field of research?

During my BSc studies, I did an elective in computational linguistics. This opened to me the wonderful and challenging world of language and natural language processing (NLP). Language poses so many challenges (way more than we humans would ever think). To equip computers with language capabilities fascinated me from the very first moment. I am very happy to contribute to this exciting and fast-growing research field whose output has potential for immediate impact on society, for example, via machine translation or digital assistants.

What are the scientific challenges and perspectives in your project?

The lack of dedicated resources and the variety of language makes Natural Language Understanding very challenging. Language is complex and constantly evolving. The project addresses one of the most important challenges in the field: improving the ability to generalize across tasks, languages and language varieties. Modern neural network have led to enormous progress in the field, but they are very data intense which leads to significant drops in accuracy. One of the project’s key challenges is to develop algorithms that learn under minimal guidance and continuously improve. Achieving this goal means opening up technology to more languages and language varieties.

What is your estimate of the impact, which your project may have to society in the long term?

My project will contribute to make Natural Language Understanding, in particular, information extraction technology from text, more accurate and make it available to more languages and language varieties. This will support information access to big multilingual document collections. As specific use case, we collaborate with Styrelsen for Arbejdsmarket en Rekruttering to analyze job skill demands in Denmark.

Which impact do you expect the Sapere Aude programme will have on your career as a researcher?

The Sapere Aude project strengthens innovation in artificial intelligence technology for text understanding. It enables me to build a leading independent research group to strengthen AI research in Denmark and the international level. It facilitates collaboration with the national public sector and the leading international research groups in the field.

Background and personal life

I am living with my husband in Copenhagen. I come from a German-speaking minority from a multilingual province in Northern Italy and I love being out in nature. The best vacation is hiking back home in the Dolomite mountains.