Liesbeth Augustinus awarded the 2016 Steven Krauwer Award for Young Scientists

Submitted by Linda Stokman on

The 2016 Steven Krauwer Award for Young Scientists is awarded to Liesbeth Augustinus (KU Leuven) for her outstanding work on making example-based treebank search in multiple languages available to linguists and other humanities scholars, for promoting this approach and training intended users, and for applying the relevant applications in her own research.

Motivation of the jury

Example-based Search as implemented in GrETEL (Greedy Extraction of Trees for Empirical Linguistics)and its successors offers the possibility to query treebanks on the basis of natural language examples. The user need not know a formal query language, and even more importantly, it significantly reduces the need for having exact knowledge of the tree structures and the grammar behind the annotations in a particular treebank.

The user provides an example utterance illustrating the construction to be searched for, and indicates which aspects of this example utterance are essential. The system then automatically generates a query in a formal query language (XPATH), and applies it to the treebank. The user can, if (s)he wishes so, see and even adapt the XPATH query. This feature is of great educational value and reduces some of the limitations of purely example-based search.

Liesbeth made GrETEL initially as a CLARIN-tool for search in 1 million token size Dutch corpora. It was extended later by her and her colleagues to search in the 500 million token SONAR corpus. It was further extended to other languages, in particular Afrikaans and English. The Poly-GrETEL variant enables crosslingual example-based querying, which is of special interest for researchers in comparative linguistics and translation studies. Other researchers are extending GrETEL in yet other ways. GrETEL is also one of the corpus search engines that is used in TaalPortaal to offer corpus search options for descriptions of grammatical constructions of Dutch.

Liesbeth actively promoted the use of GrETEL and its successors through lectures and publications,trained the intended users of the applications through workshops and summer schools, and wrote an educational module so that the use of GrETEL can also be made part of the courses given by other teachers.

Finally, Liesbeth used GrETEL in her own research, which resulted in a novel analysis of verb clusters in the framework of HPSG, written down in her PhD thesis, which was successfully defended on October 8th, 2015. This PhD thesis was, to our knowledge, the first PhD thesis based on research that crucially used CLARIN tools. She will continue her research in a new project, which will focus on the use of (parallel) treebanks for comparative syntax on Dutch, German and Afrikaans.

In short, the example-based query engine GrETEL and its successors are very suitable for humanities researchers in general and linguists in particular. It can be used for treebanks of considerable size and in several languages, even in parallel. It allows the users to start in a very simple, user-friendly way, not having to deal with a formal query language at all. When they are up to it, they can go one step further by adapting an automatically generated query. In all these respects, GrETEL is completely in line with the main CLARIN goal of applying the results of speech and language technology to support research in the humanities.

The award ceremony  

The 3rd edition of the award ceremony took place at the annual CLARIN Annual Conference in Aix-en-Provence, France.

The Steven Krauwer Award for Young Scientist Award was initiated in 2014 and is awarded yearly to an outstanding young scientist or engineer for major contributions to building language resources, developing tools and sharing knowledge. Apart from the honour the prize includes a certificate and a cash amount of 500 euro.