Analyzing and answering questions
A computer-implemented method includes receiving, at a computer system, a question; identifying one or more first semantic elements in the question; selecting, from one or more electronic documents, a plurality of candidate responses to the question based on comparison of the one or more first semantic elements to second semantic elements; determining completeness scores for the plurality of candidate responses, wherein each of the completeness scores indicates how completely a corresponding candidate response from the plurality of candidate responses answers the question; determining relevance scores for the plurality of candidate responses, wherein each of the relevance scores indicates how relevant a corresponding candidate response from the plurality of candidate responses is to the question; and providing, by the computer system, at least a portion of the plurality of candidate responses based, at least in part, on the completeness scores and the relevance scores.
1. A computer-implemented method comprising:
receiving, at a computer system, a question from a user;
identifying, by the computer system, a topic of the question;
classifying, by the computer system, the user as belonging to a particular user category of a plurality of user categories;
accessing, by the computer system, a model for the particular user category of the user, the model identifying topics for which users in the particular user category are recognized as having specialized knowledge;
identifying, by the computer system, a candidate response to the question and a level of background information that the candidate response contains on the topic of the question;
evaluating, by the computer system, the model to determine a level of pre-existing knowledge of the user with respect to the topic of the question;
comparing, by the computer system, the level of background information on the topic of the question contained in the candidate response to the level of pre-existing knowledge of the user with respect to the topic;
determining a relevance score for the candidate response, wherein the relevance score indicates how relevant the candidate response is to the question;
determining, by the computer system and based at least on the relevance score for the candidate response and a result of comparing the level of background information on the topic of the question contained in the candidate response to the level of pre-existing knowledge of the user with respect to the topic, whether to return the candidate response for presentation to the user; and
returning the candidate response for presentation to the user.
2. The computer-implemented method of claim 1 , wherein determining the candidate response to the question comprises selecting the candidate response from one or more electronic documents.
3. The computer-implemented method of claim 1 , wherein determining the candidate response comprises comparing one or more first semantic elements in the question to pre-defined semantic elements.
4. The computer-implemented method of claim 1 , wherein classifying the user as belonging to the particular user category of the plurality of user categories comprises classifying the user as having a particular organizational role or title of a plurality of pre-defined organizational roles or titles.
5. The computer-implemented method of claim 4 , comprising determining whether to return the candidate response for presentation to the user based on whether the level of background information on the topic of the question contained in the candidate response corresponds to the level of pre-existing knowledge attributed to users having the particular organizational role or title to which the user was classified.
6. The computer-implemented method of claim 1 , wherein the candidate response contains a reduced level of background information on the topic relative to other available candidate responses, and the computer system determines to return the candidate response for presentation to the user based on the user having at least a minimum level of pre-existing knowledge related to the topic of the question.
7. The computer-implemented method of claim 1 , further comprising determining a completeness score for the candidate response, wherein the completeness score indicates how completely the candidate response answers the question, wherein the computer system determines whether to return the candidate response for presentation to the user further based on the completeness score for the candidate response.
8. The computer-implemented method of claim 1 , wherein determining whether to return the candidate response for presentation to the user comprises determining a first set of semantic elements in the question from the user and comparing the first set of semantic elements to a second set of semantic elements that occur in the candidate response.
9. A system comprising:
one or more processors; and
one or more non-transitory computer-readable media having instructions stored thereon that, when executed by the one or more processors, cause performance of operations comprising:
receiving, at the system, a question from a user;
identifying, by the system, a topic of the question;
classifying, by the system, the user as belonging to a particular user category of a plurality of user categories;
accessing, by the system, a model for the particular user category of the user, the model identifying topics for which users in the particular user category are recognized as having specialized knowledge;
identifying, by the system, a candidate response to the question and a level of background information that the candidate response contains on the topic of the question;
evaluating, by the system, the model to determine a level of pre-existing knowledge of the user with respect to the topic of the question;
comparing, by the system, the level of background information on the topic of the question contained in the candidate response to the level of pre-existing knowledge of the user with respect to the topic;
determining a relevance score for the candidate response, wherein the relevance score indicates how relevant the candidate response is to the question;
determining, by the system and based at least on the relevance score for the candidate response and a result of comparing the level of background information on the topic of the question contained in the candidate response to the level of pre-existing knowledge of the user with respect to the topic, whether to return the candidate response for presentation to the user; and
returning the candidate response for presentation to the user.
10. The system of claim 9 , wherein determining the candidate response to the question comprises selecting the candidate response from one or more electronic documents.
11. The system of claim 9 , wherein determining the candidate response comprises comparing one or more first semantic elements in the question to pre-defined semantic elements.
12. The system of claim 9 , wherein classifying the user as belonging to the particular user category of the plurality of user categories comprises classifying the user as having a particular organizational role or title of a plurality of pre-defined organizational roles or titles.
13. The system of claim 9 , wherein the candidate response contains a reduced level of background information on the topic relative to other available candidate responses, and the computer system determines to return the candidate response for presentation to the user based on the user having at least a minimum level of pre-existing knowledge related to the topic of the question.
14. The system of claim 9 , wherein the operations further comprise determining a completeness score for the candidate response, wherein the completeness score indicates how completely the candidate response answers the question, wherein the system determines whether to return the candidate response for presentation to the user further based on the completeness score for the candidate response.