Machine reading comprehension system for answering queries related to a document

A system, comprising:

• at least one processor;
• and
• memory storing instructions that, when executed by the at least one processor, cause the system to perform a set of operations, the set of operations comprising:
• processing a plurality of sections of a document to generate, for each section of the plurality of sections:
• a first score for the section in relation to a received query according to a first process;
• and
• a second score for the section in relation to the received query according to a second process different from the first process;
• ranking the plurality of sections based on the respective first similarity score and the respective second similarity score for each section;
• and
• processing at least one highest-ranked section to generate a predicted answer to the received query.

A method, comprising:

• processing a plurality of sections of a document to generate, for each section of the plurality of sections:
• a first score for the section in relation to a received query according to a first process;
• and
• a second score for the section in relation to the received query according to a second process different from the first process;
• ranking the plurality of sections based on the respective first similarity score and the respective second similarity score for each section;
• processing at least one highest-ranked section to generate a predicted answer to the received query, the processing comprising:
• generating an encoded representation of the received query and an encoded representation of the at least one highest-ranked section;
• and
• generating, based on the encoded representation of the received query and the encoded representation of the at least one highest-ranked section, a predicted answer to the received query;
• and
• providing the predicted answer in response to the received query.

A method, comprising:

• processing a plurality of sections of a document to generate, for each section of the plurality of sections:
• a first score for the section in relation to a received query according to a first process;
• and
• a second score for the section in relation to the received query according to a second process different from the first process;
• ranking the plurality of sections based on the respective first similarity score and the respective second similarity score for each section;
• and
• processing at least one highest-ranked section to generate a predicted answer to the received query.