OLLIE: On-Line Learning For Information Extraction

This paper reports work aimed at developing an open, distributed learning environment, OLLIE, where researchers can experiment with different Machine Learning (ML) methods for Information Extraction. Once the required level of performance is reached, the ML algorithms can be used to speed up the manual annotation process. OLLIE uses a browser client while data storage and ML training is performed on servers. The different ML algorithms use a unified programming interface; the integration of new ones is straightforward.