Ranking passages by merging features from factoid answers

A computer program product, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform a method comprising:

• generating, by a processor, a plurality of answer-passage pairs, each associated with a respective one of a plurality of feature vectors;
• for each answer in the answer-passage pairs, merging, by the processor, the feature vectors associated with any of the answer-passage pairs that include the answer to obtain a merged answer vector for the answer;
• for each passage in the answer-passage pairs, (i) merging, by the processor, the feature vectors associated with any of the answer-passage pairs that includes the passage to obtain a merged passage vector for the passage, (ii) merging, by the processor, the feature vectors from the merged answer vector for each answer that is associated with the passage via at least one of the answer-passage pairs to obtain a merged passage-answer vector for the passage, and (iii) concatenating, by the processor, the merged passage vector and the merged passage-answer vector to obtain a concatenated passage vector for the passage;
• and
• ranking, by the processor, the concatenated passage vector for the each passage to obtain a ranked list of passages with associated confidence scores.

A computer processing system, comprising:

• a processor, configured to
• generate a plurality of answer-passage pairs, each associated with a respective one of a plurality of feature vectors;
• for each answer in the answer-passage pairs, merge the feature vectors associated with any of the answer-passage pairs that include the answer to obtain a merged answer vector for the answer;
• for each passage in the answer-passage pairs, (i) merge the feature vectors associated with any of the answer-passage pairs that includes the passage to obtain a merged passage vector for the passage, (ii) merge the feature vectors from the merged answer vector for each answer that is associated with the passage via at least one of the answer-passage pairs to obtain a merged passage-answer vector for the passage, and (iii) concatenate the merged passage vector and the merged passage-answer vector to obtain a concatenated passage vector for the passage;
• and
• rank the concatenated passage vector for the each passage to obtain a ranked list of passages with associated confidence scores.