Current question answering systems often focus on providing a simple entity or short sentence as an answer. By gaining confidence in information retrieval systems, users start to ask more complex questions that require sophisticated answers, such as reasoning chains. However, no research has been carried out yet to determine how exhaustive an answer should be. We combine Bloom’s learner’s levels of understanding with question difficulty classification. Therefore, we heuristically determine the threshold within Bloom’s taxonomy that separates question types into simple and complex. Moreover, we extract keywords from the taxonomy within question datasets and categorize questions accordingly. Then, we train a word n-gram multi-layer perceptron (MLP) and an LSTM model with syntactic features. The results are further improved by applying a genetic algorithm for parameter tuning.     «

Current question answering systems often focus on providing a simple entity or short sentence as an answer. By gaining confidence in information retrieval systems, users start to ask more complex questions that require sophisticated answers, such as reasoning chains. However, no research has been carried out yet to determine how exhaustive an answer should be. We combine Bloom’s learner’s levels of understanding with question difficulty classification. Therefore, we heuristically determine the t...     »