17
October 10, 2005
DUC 2005 @ Vancouver, Canada
Conclusions:
    A simple system features
*Concept Link: new way to calculate sentence similarity;
nno chunker/parser involved
nconcept differs from NPs in Lexical Chain
*Considering sentence similarity/relatedness via Concept Link:
nAlleviate the influence of expression variations; (but might involve inaccurate sense guess)
nOutperforms Word co-occurrence approach
*Minimizing Redundancy via Modified MMR;
*No extra heuristics involved.
Document boundaries are ignored. (No sentence position)
No grammar needed: implies the system is far from understanding the content accuractly.