Relation Extraction for Inferring Access Control Rules from Natural Language Artifacts

Presentation pdf 711KB

We propose an approach to define access control rules (ACRs) through extracting relations (i.e., the relationship among two or more items) from NL artifacts. We found ACRs exist in 47% of the evaluated sentences. We identified those sentences with an F1 performance of 72% and extracted ACRs with an F1 performance of 60%.

Author(s):

John Slankas    
North Carolina State University
United States

Xusheng Xiao    
North Carolina State University
United States

Laurie Williams    
North Carolina State University
United States

Tao Xie    
University of Illinois, Urbana-Champaign
United States