SINDI-WALKS: Workbench for PLOT-based Technological Information Extraction and Management

This paper introduces a workbench called SINDI-WALKS capable of extracting inherent valuable information from scientific literatures such as articles and patents. The system identifies PLOTs, Person, Location, Organization and Terms and extracts the semantic relations between them. Also, the workbench provides two test beds for monitoring the run-time behavior of the core engines, SINDI-CORE (PLOT recognition engine) and SINDI-LINK (relation extraction engine). As an important auxiliary tool, an intuitive annotation system is additionally furnished to construct training or evaluation collections for the engines above. Various administrative functions are also provided to visually manage the extracted semantic triples in an efficient manner. The proposed toolsets in this paper will facilitate the performance optimization of the information extraction as well as the efficient management of the extracted semantic triples.