Question answering is widespread and a vari- ety of answer taxonomies exists in research that divides responses into simple and com- plex. Multi-hop answering has become pop- ular when the complexity of questions and an- swers increases. However, determining when multi-hop reasoning becomes necessary is not yet clear. We propose to apply Bloom’s taxonomy to the determination of question complexity in question-answering systems. Originating in pedagogy, Bloom’s taxonomy measures ques- tion complexity to determine learning progress levels. Subsequently, the determined question complexity can help in deciding whether an entity or phrase is sufficient as an answer or whether reasoning chains should be given.    «

Question answering is widespread and a vari- ety of answer taxonomies exists in research that divides responses into simple and com- plex. Multi-hop answering has become pop- ular when the complexity of questions and an- swers increases. However, determining when multi-hop reasoning becomes necessary is not yet clear. We propose to apply Bloom’s taxonomy to the determination of question complexity in question-answering systems. Originating in pedagogy, Bloom’s taxonomy measures ques- tion comp...    »