Sabine Geldof

Sabine's fascination with language emerged in secondary school, where she pursued Latin and Greek studies. At Katholieke Universiteit Leuven, she earned a B.A. in Romance Philology, graduating magna cum laude with a thesis on Computational Linguistics. After gaining several years' experience in the IT industry, Sabine returned to her interest in language research.

At the Artificial Intelligence laboratory, Vrije Universiteit Brussel, she conducted extensive research on knowledge engineering and natural language generation, work that contributed to her Ph.D. from the University of Antwerp. Sabine has published widely on a range of topics, from context modelling and processing to mobile and adaptive interfaces. When work and her three children aren't clamouring for her attention, Sabine loves to travel and absorb other cultures, preferring family hiking and camping expeditions as a way of exploring a country.


What attracted you to Namahn?

I had been familiar with Namahn's work and realised there were strong links between its service offerings and my research interests. And I had worked on the automatic generation of instructions and advisory texts, which is related to technical communication.

Why did you leave Romances Languages for the field of Computational Linguistics?

I realised I was also drawn to learning about the mechanisms that underlie human language processing. Computational Linguistics is a relatively new field that combines the expertise of many disciplines like the humanities, the natural and behavioural sciences and engineering. I was intrigued by the questions that were being explored in that area. They were looking at things like how meaning is conveyed through a language's structure and what the link is between the language we use to express ourselves and how we acquire knowledge.The study of natural language generation is one of your main areas of research.

Why is it so important?

By modelling the mechanisms of language processing on a computer we can discover and articulate properties that are relevant for any type of language interaction. For instance, although the Internet combines various media — text, graphics, sound and movies — that universe is most naturally accessed through language. For things like web browsing, filtering and processing of information you need software to reach the content. Language technology also helps to manage content to make digital information available to everyone.

Describe some of your achievements at the Vrije Universiteit Brussel's Artificial Intelligence Laboratory.

I belonged to a research group that built intelligent systems. I helped develop and apply a methodology for building expert systems; modelling the knowledge that is needed for the system. I was also part of a team that worked on intelligent interfaces, where I brought in my knowledge of linguistic interaction. We developed adaptive web-based interfaces. As you browsed through the website there was a 'softbot' that helped you navigate. Softbots are intelligent agents that use software tools and series on your behalf. They dynamically determine how and where to satisfy your interests.

Did your Ph.D. thesis grow out of the work you were doing in artificial intelligence?

Yes, my dissertation research involved modelling and processing the user's context — both physical and linguistic. Our project team built a portable digital assistance device that would give you advice in an information-rich situation. The domain we applied the device to was a conference situation. When you go to a conference there are multiple activities going on. There may be many exhibits and parallel sessions taking place — some people you know and some you don't. You'd have this personal digital assistant, linked to a system with softbots that would remind you that there was a relevant presentation taking place at such and such a time. I researched how to make these messages context-sensitive, how to formulate the message, model the context and how that related to the parameters of language formulation.

How did you end up 'down under' for your post doc?

The head of the Centre for Language Technology at Sydney's Macquarie University is an authority in the field of natural language generation so I was very honoured to obtain a fellowship there. My work involved conducting research on natural language generation technology for mobile devices in the application area of route descriptions. They wanted to figure out how to improve the interface of a way-finding system, how language can improve communication with computers. Australia was fascinating on a personal level as well. It offered such a dramatic contrast to everything that had been familiar to me in Europe. For example, the difference in the light there is reflected in their distinctive way of painting. And the mood is different there because of the sense you have of being on an island. There are these archaic, primordial plants and animals in Australia that you find nowhere else in the world, that are quite extraordinary.