Frank Schilder
Research Director
Research & Development, Thomson Reuters
Eagan, Minnesota


Bio:

I am a Research Director in Thomson Reuters Research & Development
group where I lead a team of five researchers generally concentrating
on Natural Language Processing and Large-Scale Analytics. Since
joining Thomson Reuters in 2004, I have been conducting applied
research projects focused on developing summarization technologies and
information extraction systems; for example, my summarization work has
been implemented as the snippet generator for search results of
WestLawNext, a search engine for legal professionals.  I am currently
leading a Natural Language Generation project developing methods for
learning templates used for stock market reports and biographies. We
published our work on the GenNext system at different conferences
(IWCS, ACL, ENLG) last year.

In the past, I have successfully participated in several research
competitions on automatic summarization systems such as the Text
Analysis Conference (TAC) carried out by the National Institute of
Standards and Technology (NIST).  I have served on several program
committees and co-organized a Dagstuhl workshop on Annotating,
Extracting and Reasoning about Time and Events.  I co-authored (with
Peter Jackson) an overview chapter on NLP in the Encyclopedia of
Language & Linguistics and published a chapter on multi-lingual
summarization (together with Liang Zhou). Before joining Thomson
Reuters, I was employed by the Department for Informatics at the
University of Hamburg, Germany, as an assistant professor.


Candidate Statement:

Natural Language Generation has received an increased interest in
recent years in the publishing industry that makes it ripe for
interesting applications in various business segments such as news,
but also the legal and the scientific communities. In particular, I
have seen first hand that personalization of news and automatic report
generations including visualizations are an increasingly important
part of NLG. This community has a unique opportunity to contribute to
this development.

If elected, I will promote shared tasks based on real-world
applications, work to obtain data sets for the NLG community, and
encourage cross-fertilization between the summarization and NLG
communities.