Interpretation and normalization of temporal expressions for question answering

The German question answering (QA) system InSicht participated in QA@CLEF for the third time. InSicht implements a deep QA approach: it builds on full sentence parses, inferences on semantic representations. and matching semantic representations derived from questions and documents. InSicht was improved for QA@CLEF 2006 as follows: temporal expressions are normalized and temporal deictic expressions are resolved to explicit date representations;, the coreference module was extended by a fallback strategy for increased robustness;, equivalence rules can introduce negated relations;, answer candidates are clustered in order to avoid multiple occurrences of one real-world entity in the answers to a list question; and finally a shallow QA subsystem that produces a second answer stream was integrated into InSicht. The current system is evaluated in an ablation study on the German questions from QA@CLEF 2006.