A mechanism is provided in a data processing system comprising at least one processor and a memory comprising instructions which, when executed by the at least one processor, causes the at least one processor to train a similar passage cognitive system. The mechanism receives a question and answer key for a question answering cognitive system, the question and answer key comprising a list of question and answer specification pairs. Each question is a text string and each answer specification references one or more text passages from a corpus of information. The mechanism uses the question and answer key to generate a similar passage map for the similar passage cognitive system, the similar passage map comprising a list of text relation pairs. Each text relation pair comprises a sample input text component and a list comprising one or more sample output text components. The mechanism trains a similar passage machine learning model of the similar passage cognitive system using the similar passage map.