A Sentence Simplification System for Improving Relation Extraction
This work addresses a specific bottleneck for researchers and practitioners in natural language processing by enhancing Open RE accuracy through simplification, though it is incremental as it builds on existing methods.
The paper tackles the problem of syntactically complex sentences hindering Open Relation Extraction (RE) systems by developing a text simplification framework that pre-processes sentences using syntactic transformation rules, resulting in improved performance for state-of-the-art Open RE systems.
In this demo paper, we present a text simplification approach that is directed at improving the performance of state-of-the-art Open Relation Extraction (RE) systems. As syntactically complex sentences often pose a challenge for current Open RE approaches, we have developed a simplification framework that performs a pre-processing step by taking a single sentence as input and using a set of syntactic-based transformation rules to create a textual input that is easier to process for subsequently applied Open RE systems.