Teaching Machines to Read and Comprehend

NeurIPS 2015 Karl Moritz HermannTomáš KočiskýEdward GrefenstetteLasse EspeholtWill KayMustafa SuleymanPhil Blunsom

Teaching machines to read natural language documents remains an elusive challenge. Machine reading systems can be tested on their ability to answer questions posed on the contents of documents that they have seen, but until now large scale training and test datasets have been missing for this type of evaluation... (read more)

PDF Abstract

Evaluation results from the paper


Task Dataset Model Metric name Metric value Global rank Compare
Question Answering CNN / Daily Mail Impatient Reader CNN 63.8 # 15
Question Answering CNN / Daily Mail Impatient Reader Daily Mail 68.0 # 10
Question Answering CNN / Daily Mail Attentive Reader CNN 63 # 16
Question Answering CNN / Daily Mail Attentive Reader Daily Mail 69 # 8
Question Answering CNN / Daily Mail MemNNs (ensemble) CNN 69.4 # 13