Cross-lingual German Biomedical Information Extraction: from Zero-shot to Human-in-the-Loop
This addresses the problem of data scarcity in German biomedical text processing, but it is incremental as it builds on existing methods without introducing new paradigms.
The paper tackles biomedical information extraction from German clinical narratives with limited annotations, proposing a project that combines transfer learning and active learning strategies along with a user interface for model inspection and annotation.
This paper presents our project proposal for extracting biomedical information from German clinical narratives with limited amounts of annotations. We first describe the applied strategies in transfer learning and active learning for solving our problem. After that, we discuss the design of the user interface for both supplying model inspection and obtaining user annotations in the interactive environment.