My long-term research goal is to make
computers achieve human levels of performance when using natural language.
More precisely, I am interested in developing
theories and algorithms that can lead to
computer software capable of producing meaningful sentences and structured,
coherent text from bits and pieces of textual information.
These capabilities, deployed in applications such as translation,
summarization, and question answering,
could significantly increase our ability to assimilate and
disseminate information.
I have got my Ph.D. in August 2006 with a dissertation entitled
"Natural Language Generation
using an Information-Slim Representation", under the
supervision of Prof. Daniel Marcu.
In the Summer of 2003 I worked with Eric Brill in
the Machine Learning
and Applied Statistics group at Microsoft Research on
question answering and automatic evaluation.
Radu Soricut. 1996.
Ring-buffer Protocols in Process Algebra.
Proceedings "Tendinte Actuale in Cibernetica si Filosofia
Sttintei", National Conference of the Academy of Cybernetics "Stefan
Odobleja". Baile Felix, Romania, October 1996.