Candidate answers for speculative questions in a deep question answering system

Abstract

Techniques for generating answers to speculative questions are provided. It is determined that a question received by a deep question answering system is speculative, based at least in part on identifying a future tense in the question. A set of candidate answers is generated by one or more predictive algorithms, where each of the one or more predictive algorithms are used to generate at least one estimated future value for an attribute of the question. A score is computed for each candidate answer in the set of candidate answers, and a first candidate answer, of the set of candidate answers, is returned as responsive to the speculative question received by the deep question answering system.

What is claimed is: 
     
         1 . A method, comprising:
 determining that a question received by a deep question answering system is speculative, based at least in part on identifying a future tense in the question;   generating, by one or more predictive algorithms, a set of candidate answers, wherein each of the one or more predictive algorithms are used to generate at least one estimated future value for an attribute of the question;   computing a score for each candidate answer in the set of candidate answers; and   returning a first candidate answer, of the set of candidate answers, as responsive to the speculative question received by the deep question answering system.   
     
     
         2 . The method of  claim 1 , further comprising:
 storing an indication specifying to verify each of the candidate answers subsequent to a time related to the question; and   upon determining that the time has passed:
 comparing each candidate answer in the set of candidate answers to an actual answer to the question to determine an accuracy of each candidate answer; 
 adjusting a confidence score of an ability of the deep question answering system to generate the set of candidate answers; and 
 adjusting a confidence score of the predictive algorithm that generated the candidate answer in the set of candidate answers based on the accuracy of each respective candidate answer. 
   
     
     
         3 . The method of  claim 1 , wherein determining that the question is speculative comprises at least one of: (i) identifying the future tense in the question, (ii) identifying an independent variable in the question, and (iii) determining that a corpus of information does not include a candidate answer for the question. 
     
     
         4 . The method of  claim 1 , wherein generating the set of candidate answers comprises:
 collecting relevant evidence from a corpus of information;   selecting the one or more predictive algorithms; and   applying the one or more predictive algorithms to the relevant evidence and at least one attribute of the question.   
     
     
         5 . The method of  claim 1 , further comprising:
 computing a confidence score of an analysis of the question and supporting evidence gathered from a corpus of information based on the analysis of the question; and   adjusting the score of each candidate answer based on the computed confidence score.   
     
     
         6 . The method of  claim 1 , wherein the one or more predictive algorithms are selected from: (i) a set of existing predictive algorithms, and (ii) one or more predictive algorithms generated by the deep question answering system. 
     
     
         7 . The method of  claim 6 , wherein the predictive algorithms generated by the deep question answering system are based on: (i) at least one attribute of the question, (ii) relevant evidence gathered from a corpus of information, (iii) one or more trends found in the corpus of information. 
     
     
         8 . A system, comprising:
 one or more computer processors; and   a memory containing a program which when executed by the one or more computer processors performs an operation, the operation comprising:
 determining that a question received by a deep question answering system is speculative, based at least in part on identifying a future tense in the question; 
 generating, by one or more predictive algorithms, a set of candidate answers, wherein each of the one or more predictive algorithms are used to generate at least one estimated future value for an attribute of the question; 
 computing a score for each candidate answer in the set of candidate answers; and 
 returning a first candidate answer, of the set of candidate answers, as responsive to the speculative question received by the deep question answering system. 
   
     
     
         9 . The system of  claim 8 , the operation further comprising:
 storing an indication specifying to verify each of the candidate answers subsequent to a time related to the question; and   upon determining that the time has passed:
 comparing each candidate answer in the set of candidate answers to an actual answer to the question to determine an accuracy of each candidate answer; 
 adjusting a confidence score of an ability of the deep question answering system to generate the set of candidate answers; and 
 adjusting a confidence score of the predictive algorithm that generated the candidate answer in the set of candidate answers based on the accuracy of each respective candidate answer. 
   
     
     
         10 . The system of  claim 8 , wherein determining that the question is speculative comprises at least one of: (i) identifying the future tense in the question, (ii) identifying an independent variable in the question, and (iii) determining that a corpus of information does not include a candidate answer for the question. 
     
     
         11 . The system of  claim 8 , wherein generating the set of candidate answers comprises:
 collecting relevant evidence from a corpus of information;   selecting the one or more predictive algorithms; and   applying the one or more predictive algorithms to the relevant evidence and at least one attribute of the question.   
     
     
         12 . The system of  claim 8 , the operation further comprising:
 computing a confidence score of an analysis of the question and supporting evidence gathered from a corpus of information based on the analysis of the question; and   adjusting the score of each candidate answer based on the computed confidence score.   
     
     
         13 . The system of  claim 8 , wherein the one or more predictive algorithms are selected from: (i) a set of existing predictive algorithms, and (ii) one or more predictive algorithms generated by the deep question answering system. 
     
     
         14 . The system of  claim 13 , wherein the predictive algorithms generated by the deep question answering system are based on: (i) at least one attribute of the question, (ii) relevant evidence gathered from a corpus of information, (iii) one or more trends found in the corpus of information. 
     
     
         15 . A computer program product, comprising:
 a computer-readable storage medium having computer-readable program code embodied therewith, the computer-readable program code comprising:
 computer-readable program code configured to determine that a question received by a deep question answering system is speculative, based at least in part on identifying a future tense in the question; 
 computer-readable program code configured to generate, by one or more predictive algorithms, a set of candidate answers, wherein each of the one or more predictive algorithms are used to generate at least one estimated future value for an attribute of the question; 
 computer-readable program code configured to compute score for each candidate answer in the set of candidate answers; and 
 computer-readable program code configured to return a first candidate answer, of the set of candidate answers, as responsive to the speculative question received by the deep question answering system. 
   
     
     
         16 . The computer-program product of  claim 15 , further comprising:
 computer-readable program code configured to store an indication specifying to verify each of the candidate answers subsequent to a time related to the question; and   computer-readable program code configured to, upon determining that the time has passed:
 compare each candidate answer in the set of candidate answers to an actual answer to the question to determine an accuracy of each candidate answer; 
 adjust a confidence score of an ability of the deep question answering system to generate the set of candidate answers; and 
 adjust a confidence score of the predictive algorithm that generated the candidate answer in the set of candidate answers based on the accuracy of each respective candidate answer. 
   
     
     
         17 . The computer program product of  claim 15 , wherein determining that the question is speculative comprises at least one of: (i) identifying the future tense in the question, (ii) identifying an independent variable in the question, and (iii) determining that a corpus of information does not include a candidate answer for the question. 
     
     
         18 . The computer program product of  claim 15 , wherein generating the set of candidate answers comprises:
 collecting relevant evidence from a corpus of information;   selecting the one or more predictive algorithms; and   applying the one or more predictive algorithms to the relevant evidence and at least one attribute of the question.   
     
     
         19 . The computer program product of  claim 16 , further comprising:
 computer-readable program code configured to compute a confidence score of an analysis of the question and supporting evidence gathered from a corpus of information based on the analysis of the question; and   computer-readable program code configured to adjust the score of each candidate answer based on the computed confidence score.   
     
     
         20 . The computer program product of  claim 15 , wherein the one or more predictive algorithms are selected from: (i) a set of existing predictive algorithms, and (ii) one or more predictive algorithms generated by the deep question answering system, wherein the predictive algorithms generated by the deep question answering system are based on: (i) at least one attribute of the question, (ii) relevant evidence gathered from a corpus of information, (iii) one or more trends found in the corpus of information.