Lecture: 'Data-To-Text: Generating Textual Summaries of Complex Data'

xoves, 6 setembro, 2018 -
11:00 - 12:00
CiTIUS Assembly Hall
Ehud Reiter (University of Aberdeen, Scotland)

Data-to-Text systems use Natural Language Generation (NLG) technology to generate summaries in English (and other human languages) of complex non-linguistic (usually numeric) data sets. 

In my talk, I will give a general introduction to data-to-text, including examples of systems and applications, current technology and evaluations, and an overview of commercial activity in this area.  I will also briefly summarise some of the current research in my group on improving data-to-text technology.

About Prof. Reiter

Professor Ehud Reiter is Full Professor with the University of Aberdeen, Scotland. His education (BA, MA, PhD) was all at Harvard University (USA). He has both commercial and academic background. He has been a lecturer in the Aberdeen Computing Science Department since August 1995. He founded the Natural Language Generation (NLG) research group at Aberdeen. Before that he was a Senior Scientist in CoGenTex, a small (USA) NLP software house; a Research Fellow at the University of Edinburgh (UK); and a Research Associate at the University College of Cork (Ireland). He is currently Chief Scientist of Arria NLG company. His research interests include natural-language generation, data-to-text, medical informatics, multimodal systems, etc. He is world-wide recognized as pioneer in the NLG research field. The book entitled “Building natural language generation systems” (co-authored by E. Reiter and R. Dale) is considered as an essential reference for all readers in the field. According to Google Scholar (accessed on April 4, 2017) he has h-index=38 and i10-index=92.