Intelligent search engine

Abstract

A search engine ( 200, 500, 800 ), method and a system for performing a search is provided. The search engine ( 200, 500, 800 ) is connected to at least one mobile device ( 210, 510, 810 ) and at least one web crawler ( 222, 522, 822 ). The web crawler ( 222, 522, 822 ) is configured to index documents and classify said documents. The search engine ( 200, 500, 800 ) receives a query from the mobile device ( 210, 510, 810 ) which is determined to be best answered by a crowd-sourced answer. The search engine ( 200, 500, 800 ) searches the documents and delivers at least one crowd-sourced answer ( 318, 618, 918 ). The search engine ( 200, 500, 800 ) displays the crowd-sourced answer ( 318, 618, 918 ) to a user.

1 . A search engine ( 200 ,  500 ,  800 ) connected to at least one mobile device ( 210 ,  510 ,  810 ) and at least one web crawler ( 222 ,  522 ,  822 ), characterized in that,
 the web crawler ( 222 ,  522 ,  822 ) is configured to index documents and classify said documents,   the determination to seek a crowd-sourced answer is done for queries that are determined not to have a factual answer, -   the determination to seek a crowd-sourced answer is not done for queries that are determined to have a factual answer,   the factuality of the answer and/or the query is determined based on the dispersion of different answers to the query, so that a greater dispersion among different answers to the query is determined to imply non-factuality of the query and/or the answer, and a lesser dispersion of different answers to the query is determined to imply a greater factuality of the query and/or the answer.   
     
     
         2 . A search engine ( 200 ,  500 ,  800 ) as claimed in  claim 1 , characterized in that, a most popular crowd-sourced answer is subjected to a veracity test, and if the veracity test is failed, the most popular search result passing the veracity test is ranked first. 
     
     
         3 . A search engine ( 200 ,  500 ,  800 ) as claimed in  claim 1 , characterized in that, the web crawler ( 222 ,  522 ,  822 ) is configured to crawl and index any of the following individually or in a mix: text, voice, image and/or video. 
     
     
         4 . A search engine ( 200 ,  500 ,  800 ) as claimed in  claim 1 , characterized in that, a crowd-sourced answer may be sought to a query that is contextual, and/or context data required to answer the query is derived from the mobile device ( 210 ,  510 ,  810 ) of the user. 
     
     
         5 . A search engine ( 200 ,  500 ,  800 ) as claimed in  claim 4 , characterized in that, a most popular crowd-sourced answer is calculated by assigning different context weights to different results. 
     
     
         6 . A search engine ( 200 ,  500 ,  800 ) as claimed in  claim 1 , characterized in that, the search engine ( 200 ,  500 ,  800 ) is trained with a training set of queries and a validation set of queries. 
     
     
         7 . A search engine ( 200 ,  500 ,  800 ) as claimed in  claim 1 , characterized in that, the search engine ( 200 ,  500 ,  800 ) is trained with a training set of web crawler and/or index syntaxes and a validation set of web crawler and/or index syntaxes. 
     
     
         8 . A method of performing a search by a search engine, the search engine ( 200 ,  500 ,  800 ) connected to at least one mobile device ( 210 ,  510 ,  810 ) and at least one web crawler ( 222 ,  522 ,  822 ), characterized in that,
 configuring the web crawler ( 222 ,  522 ,  822 ) to index documents and classify said documents,   determining to seek a crowd-sourced answer for queries that are determined not to have an unambiguous factual answer,   determining to seek a crowd-sourced answer is not done for queries that are determined to have an unambiguous factual answer,   the factuality of the answer and/or the query is determined based on the dispersion of different answers to the query, so that a greater dispersion among different answers to the query is determined to imply non-factuality of the query and/or the answer, and a lesser dispersion of different answers to the query is determined to imply a greater factuality of the query and/or the answer.   
     
     
         9 . The method as claimed in  claim 8 , characterized in that, subjecting the most popular crowd-sourced answer to a veracity test, and if the veracity test is failed, the most popular search result passing the veracity test is ranked first. 
     
     
         10 . The method as claimed in  claim 8 , characterized in that, the web crawler ( 222 ,  522 ,  822 ) crawls and indexes any of the following individually or in a mix: text, voice, image and/or video. 
     
     
         11 . The method as claimed in  claim 8 , characterized in that, seeking a crowd-sourced answer to a query that is contextual, and/or context data required to answer the query is derived from the mobile device ( 210 ,  510 ,  810 ) of the user. 
     
     
         12 . The method as claimed in  claim 11 , characterized in that, calculating the most popular crowd-sourced answer by assigning different context weights to different results. 
     
     
         13 . The method as claimed in  claim 8 , characterized in that, training the search engine ( 200 ,  500 ,  800 ) with a training set of queries and a validation set of queries. 
     
     
         14 . The method as claimed in  claim 8 , characterized in that, training the search engine ( 200 ,  500 ,  800 ) with a training set of web crawler and/or index syntaxes and a validation set of web crawler and/or index syntaxes. 
     
     
         15 . A system for performing a search through a search engine, the search engine ( 200 ,  500 ,  800 ) connected to at least one mobile device ( 210 ,  510 ,  810 ) and at least one web crawler ( 222 ,  522 ,  822 ), characterized in that,
 a configuration module ( 212 ,  512 ) of the search engine ( 200 ,  500 ,  800 ) configures the web crawler ( 222 ,  522 ,  822 ) to index documents and classify said documents,   the AI module ( 216 ,  516 ,  816 ) is configured to seek a crowd-sourced answer for queries that are determined not to have an unambiguous factual answer,   the AI module ( 216 ,  516 ,  816 ) is configured to not seek a crowd-sourced answer for queries that are determined to have an unambiguous factual answer,   the factuality of the answer and/or the query is determined based on the dispersion of different answers to the query, so that a greater dispersion among different answers to the query is determined to imply non-factuality of the query and/or the answer, and a lesser dispersion of different answers to the query is determined to imply a greater factuality of the query and/or the answer.   
     
     
         16 . The system as claimed in  claim 15 , characterized in that, a veracity module ( 532 ) configured to subject the most popular crowd-sourced answer to a veracity test, and if the veracity test fails, a most popular search result that passes the veracity test is ranked first. 
     
     
         17 . The system as claimed in  claim 15 , characterized in that, the web crawler ( 222 ,  522 ,  822 ) is configured to crawl and index any of the following individually or in a mix: text, voice, image and/or video. 
     
     
         18 . The system as claimed in  claim 15 , characterized in that, a determination module ( 852 ) configured to seek a crowd-sourced answer to a query that is contextual and/or a context module derives context data required to answer the query from the mobile device ( 210 ,  510 ,  810 ) of the user. 
     
     
         19 . The system as claimed in  claim 18 , characterized in that, a calculation module ( 854 ) is configured to calculate the most popular crowd-sourced answer by assigning different context weights to different results. 
     
     
         20 . The system as claimed in  claim 15 , characterized in that, a training module ( 856 ) is configured to train the search engine ( 200 ,  500 ,  800 ) with a training set of queries and a validation set of queries. 
     
     
         21 . The system as claimed in  claim 15 , characterized in that, the training module ( 856 ) is configured to train the search engine ( 200 ,  500 ,  800 ) with a training set of web crawler and/or index syntaxes and a validation set of web crawler and/or index syntaxes. 
     
     
         22 . A search engine ( 200 ,  500 ,  800 ) as claimed in  claim 5 , characterized in that, the said context weights are user location dependent and/or user time dependent. 
     
     
         23 . The method as claimed in  claim 12 , characterized in that, the said context weights are user location dependent and/or user time dependent. 
     
     
         24 . The system as claimed in  claim 19 , characterized in that, the said context weights are user location dependent and/or user time dependent.