US 12,093,657 B2
Computer assisted answering Boolean questions with evidence
Sara Rosenthal, Spring Valley, NY (US); Jeffrey Scott McCarley, Bedford Hills, NY (US); and Mihaela Ancuta Bornea, White Plains, NY (US)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Nov. 5, 2021, as Appl. No. 17/519,911.
Prior Publication US 2023/0148228 A1, May 11, 2023
Int. Cl. G06F 40/40 (2020.01); G06F 40/20 (2020.01); G06F 40/30 (2020.01)
CPC G06F 40/40 (2020.01) [G06F 40/20 (2020.01); G06F 40/30 (2020.01)] 11 Claims
OG exemplary drawing
 
1. A computer-implemented method of using natural language software to respond to a Boolean question, the method comprising:
receiving a question and at least one passage;
classifying the question as a Boolean type of question;
generating evidence from the at least one passage to answer the Boolean type of question;
generating a response to the Boolean type of question using the generated evidence, wherein the response is a fine-grained Boolean type of answer labeled as a weak-No answer, a weak-Yes answer, both Yes and No answer, or a scaled type of answer; and
providing the generated evidence that supports the response to the Boolean type of question.