Fault Injection in Human-Readable Information

Abstract

An approach is provided in which a fault-injecting system injects a natural language fault into a first text segment to produce a second text segment that are both written in a natural language. The fault-injecting system receives a third text segment from a reviewer that includes at least one correction to the second text segment. The fault-injecting system compares the third text segment against the first text segment and generates an efficacy score. The efficacy score indicates whether the correction in the third text segment corrects the natural language fault. In turn, the fault-injecting system sends the efficacy score to an author of the first text segment.

1 . A method implemented by an information handling system that includes a memory and a processor, the method comprising:
 injecting a natural language fault into a first text segment to produce a second text segment, wherein the first text segment and the second text segment are written in a natural language;   receiving a third text segment from a reviewer, wherein the third text segment comprises at least one correction to the second text segment;   generating an efficacy score by comparing the third text segment against the first text segment, wherein the efficacy score indicates whether the at least one correction corrects the natural language fault; and   sending the efficacy score to an author of the first text segment.   
     
     
         2 . The method of  claim 1  further comprising:
 selecting a first fault class from a plurality of fault classes based on a first contextual meaning of the first text segment, wherein the natural language fault includes a first rule and belongs to the first fault class; 
 selecting the reviewer from a plurality of reviewers based on the selected first fault class; and 
 sending the second text segment to the reviewer to review, wherein the second text segment has a second contextual meaning that is different than the first contextual meaning. 
 
     
     
         3 . The method of  claim 2  further comprising:
 determining that the first fault class is a noun swap fault class; and 
 swapping a first noun in the first text segment with a second noun in the first text segment based on the first rule to produce the second text segment. 
 
     
     
         4 . The method of  claim 2  further comprising:
 determining that the first fault class is a factual discrepancy fault class; and 
 modifying at least one fact in the first text segment based on the first rule to produce the second text segment. 
 
     
     
         5 . The method of  claim 2  further comprising:
 determining that the first fault class is a sentiment modifier fault class; and 
 modifying a sentiment of the first text segment based on the first rule to produce the second text segment. 
 
     
     
         6 . The method of  claim 1  further comprising:
 updating a reviewer profile of the reviewer based on the efficacy score and a fault class of the natural language fault; 
 injecting a different natural language fault into a different first text segment based on the updated reviewer profile to produce a different second text segment; and 
 sending the different second text segment to the reviewer to review. 
 
     
     
         7 . The method of  claim 1  further comprising:
 injecting a different natural language fault into the first text segment to produce a different second text segment; 
 selecting a different reviewer based on the different natural language fault; and 
 sending the different second text segment to the different reviewer to review.