Extractive Question Answering

Extractive Question Answering (EQA) focuses on identifying the answer to a question within a given text by extracting the relevant span of words. Current research emphasizes improving model robustness against data biases and distribution shifts, particularly through novel training methodologies and the incorporation of unanswerable questions. Large language models (LLMs) are increasingly used, but their limitations in handling closed-domain knowledge and long answer spans are actively being addressed. EQA's impact spans various fields, including healthcare (e.g., extracting information from medical records) and finance (e.g., summarizing financial documents), where reliable and accurate information extraction is crucial.