Systems and methods for providing culturally-relevant search results to users

1. A computer-implemented method for providing search results to a user, the method comprising:

receiving a search query from the user, the search query being associated with a particular date and comprising a plurality of search terms;

forming, using a processor, a query feature vector for the search query, the query feature vector comprising a set of numerical values associated with the search terms;

accessing news feature vectors associated with documents related to current events, the current events documents having publication dates that are temporally proximate to the particular date, and the news feature vectors comprising sets of numerical values associated with terms in corresponding ones of the current events documents;

generating an augmented query feature vector based on the query feature vector and at least one of the news feature vectors, the generating comprising:

identifying a subset of the news feature vectors associated with at least one of the search terms and at least one of the terms within the current events documents;

generating a centroid feature vector for the subset of the news feature vectors, based on the sets of numerical values of the subset of the news feature vectors;

forming the augmented query feature vector, based on a comparison of the query feature vector and the centroid feature vector,

accessing target feature vectors associated with target documents, each the target feature vectors comprising sets of numerical values associated with terms in corresponding ones of the target documents;

computing first metrics of similarity between the augmented query feature vector and the target feature vectors, the first similarity metrics comprising at least one of distances or angles between the augmented query feature vector and the target feature vectors;

identifying search results based on the computed first similarity metrics, the search results comprising information associated with at least a portion of the target documents; and

enabling the user to perceive at least one of the identified search results.

2. The method of claim 1 , wherein generating the augmented query feature vector comprises:

calculating, using the set of numerical values of the query feature vector and the sets of numerical values of the news feature vectors as coordinates, second metrics of similarity between the query feature vector and the news feature vectors; and

selecting the subset of the news feature vectors, based on at the second similarity metrics.

3. The method of claim 2 , wherein calculating the second similarity metrics comprises calculating Euclidean distances between a point in n-dimensional space represented by the set of numerical values of the query feature vector and points in n-dimensional space represented by the sets of numerical values of the subset of the news feature vectors.

4. The method of claim 2 , wherein calculating the second similarity metrics comprises calculating angles between a vector that extends to a point in n-dimensional space represented by the set of numerical values of the query feature vector and vectors that extend to points in n-dimensional space represented by the sets of numerical values of the subset of the news feature vectors.

5. The method of claim 1 , wherein forming the augmented query feature vector comprises:

identifying terms within the current events documents associated with the centroid feature vector that are absent from the search query, based on at least the numerical values of the query feature vector and corresponding ones of the news feature vectors;

modifying the terms associated with the search query, the modified terms including the identified terms; and

generating, for the augmented query feature vector, a set of numerical values associated with the modified terms.

6. The method of claim 1 , wherein generating the augmented query feature vector comprises:

identifying terms within the search query that are absent from the current events documents associated with the centroid feature vector, based on at least the numerical values of the query feature vector and corresponding ones of the news feature vectors;

modifying the terms associated with the search query, the modified terms excluding the identified terms; and

generating, for the augmented query feature vector, a set of numerical values associated with the modified terms.

7. The method of claim 1 , wherein generating the augmented query feature vector comprises:

identifying terms within the search query that are similar to corresponding terms in the current events documents associated with the centroid feature vector;

modifying the terms associated with the search query, the modifying comprising replacing the identified terms within the similar terms from the current events documents associated with the centroid feature vector; and

generating, for the augmented query feature vector, numerical values associated with the modified terms.

8. The method of claim 1 , wherein computing the first similarity metrics comprises calculating Euclidean distances between a point in n-dimensional space represented by a set of numerical values representative of the augmented query feature vector and points in n-dimensional space represented by the sets of numerical values of the target feature vectors.

9. The method of claim 2 , wherein computing the first similarity metrics comprises calculating angles between a vector that extends to a point in n-dimensional space represented by a set of numerical values representative of the augmented query feature vector and vectors that extend to points in n-dimensional space represented by the sets of numerical values of the target feature vectors.

10. The method of claim 1 , wherein the set of numerical values of the query feature vector represent corresponding ones of the terms submitted in the search query.

11. The method of claim 1 , wherein:

the news feature vectors are associated with corresponding ones of the current events documents; and

the target feature vectors are associated with corresponding ones of the target documents.

12. The method of claim 1 , wherein at least one of the current events documents was published by a news outlet less than or equal to 24 hours prior to submission of the search query by the user.

13. The method of claim 1 , wherein at least one of the current events documents is associated with science and was published less than or equal to 6 months prior to submission of the search query by the user.

14. The method of claim 1 , wherein at least one of the current events documents or target documents are web pages.

15. The method of claim 1 , further comprising:

determining a relevance of the identified search results to the search query; and

sorting the identified search results based on the determined relevance.

16. The method of claim 1 , wherein generating the centroid feature vector comprises:

calculating a weighted average of the sets of numerical values that represent the subset of the news result feature vectors; and

generating numerical values for the centroid feature vector, the numerical values comprising at least one of the weighted averages.

17. The method of claim 1 , wherein generating the centroid feature vector comprises:

performing a clustering analysis of the sets of numerical values of the subset of the news feature vectors; and

generating numerical values for the centroid feature vector, based on at least the clustering analysis.

18. The method of claim 1 , further comprising:

accessing a target corpus data store that references the target documents, wherein the target documents include one or more documents related to current events; and

forming the target feature vectors for the target documents from the target corpus data store.

19. The method of claim 1 , further comprising:

accessing a news corpus data store that references the current events documents, wherein the current events are events related to human action; and

forming the news feature vectors for the current events documents from the news corpus data store.

20. The method of claim 1 , wherein enabling the user to perceive the search results comprises:

determining a relationship between the search results and a current event at least one of the current events; and

providing the search results to the user as a perceivable list of search results organized with search results that are closely related to the current event being located in positions within the list that are above other search results.

21. The method of claim 1 , wherein enabling the user to perceive the search results comprises:

distinguishing a first portion of the search results that are closely related to a current event and a second portion of the search results that have no particular known relationship with the current event; and

providing the search results to the user as a perceivable list of search results, the first search results being separated from the second search results using a label.

22. The method of claim 1 , wherein generating the augmented query feature vector further comprises:

identifying first and second subsets of the news feature vectors;

generating a first centroid feature vector and a second centroid vector, the first centroid vector being associated with the first subset of the news feature vectors, and the second centroid vector being associated with the second subset of the news feature vectors; and

forming the augmented query feature vector, based on a comparison of the query feature vector and the first and second centroid feature vectors.

23. The method of claim 22 , wherein:

the first and second subsets are associated with at least one of the search terms;

the first subset is associated with a first term in the current events documents;

the second subset is associated with a second term in the current events documents; and

the first term is different from the second term.

24. A system for providing search results to a user, the system comprising:

means for receiving a search query from the user, the search query being associated with a particular date and comprising a plurality of search terms;

means for forming, using a processor, a query feature vector for the search query, the query feature vector comprising a set of numerical values associated with the search terms;

means for accessing news feature vectors associated with documents related to current events, the current events documents having publication dates that are temporally proximate to the particular date, and the news feature vectors comprising sets of numerical values associated with terms in corresponding ones of the current events documents;

means for generating an augmented query feature vector based on the query feature vector and the news feature vectors, the means for generating comprising:

means for identifying a subset of the news feature vectors associated with at least one of the search terms and at least one of the terms in the current events documents;

means for generating a centroid feature vector for the subset of the news feature vectors, based on the sets of numerical values of the subset of the news feature vectors;

means for forming the augmented query feature vector, based on a comparison of the query feature vector and the centroid feature vector;

means for accessing target feature vectors associated with target documents, the target feature vectors comprising sets of numerical values associated with terms in corresponding ones of the target documents;

means for computing first metrics of similarity between the augmented query feature vector and the target feature vectors, the first similarity metrics comprising at least one of distances or angles between the augmented query feature vector and the target feature vectors;

means for identifying search results based on the computed first similarity metrics, the search results comprising information associated with at least a portion of the target documents; and

means for enabling the user to perceive at least one of the identified search results.

25. A computer-implemented method for providing search results to a user, the method comprising:

receiving a search query from the user, the search query being associated with a particular date and comprising a plurality of search terms;

forming, using a processor, a query feature vector for the search query, the query feature vector comprising a set of numerical values associated with the search terms;

accessing news feature vectors associated with documents related to current events, the current events documents having publication dates that are temporally proximate to the particular date, and the news feature vectors comprising sets of numerical values associated with terms in corresponding ones of the current events documents;

identifying a subset of the news feature vectors associated with at least one of the search terms and at least one of the terms in the current events documents;

generating a centroid feature vector for the subset of the news feature vectors, based on the sets of numerical values of the subset of the news feature vectors;

accessing target feature vectors associated with target documents, the target feature vectors comprising sets of numerical values associated with terms in corresponding ones of the target documents;

computing first metrics of similarity between comparing the query feature vector with and the target feature vectors, the first similarity metrics comprising at least one of distances or angles between the query feature vector and the target feature vectors;

identifying a set of target documents as search results for the search query based on the computed first similarity metrics;

computing a second metric of similarity between the query feature vector with the centroid feature vector, the second similarity metrics comprising at least one of a distance or an angle between the query feature vector and the centroid feature vector; and

generating a signal to present a visual display of the search results to the user, the visual display visually distinguishing a subset of the set of target documents from other documents in the search results based the second similarity metric.

26. The method of claim 25 , wherein computing the first similarity metrics comprises calculating Euclidean distances between a point in n-dimensional space represented by a set of numerical values representative of the query feature vector and points in n-dimensional space represented by the sets of numerical values of the target feature vectors.

27. The method of claim 25 , wherein computing the first similarity metrics comprises calculating angles between a vector that extends to a point in n-dimensional space represented by a set of numerical values representative of the query feature vector and vectors that extend to points in n-dimensional space represented by the sets of numerical values of the target feature vectors.

28. The method of claim 25 , wherein the numerical values of the query feature vector represent corresponding ones of the terms submitted in the search query.

29. The method of claim 25 , wherein:

the news feature vectors are associated with corresponding ones of the current events documents; and

the target feature vectors are associated with corresponding ones of the target documents.

30. The method of claim 25 , wherein presenting a visual display of the target documents comprises presenting a list of search results organized with search results that are closely related to the current event being located in positions within the list that are above other search results.

31. A tangible, non-transitory computer-readable medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform a method, comprising:

receiving a search query from the user, the search query being associated with a particular date and comprising a plurality of search terms;

forming, using a processor, a query feature vector for the search query, the query feature vector comprising a set of numerical values associated with the search terms;

accessing news feature vectors associated with documents related to current events, the current events documents having publication dates that are temporally proximate to the particular date, and the news feature vectors comprising sets of numerical values associated with terms in corresponding ones of the current events documents;

generating an augmented query feature vector based on the query feature vector and at least one of the news feature vectors, the generating comprising:

identifying a subset of the news feature vectors associated with at least one of the search terms and at least one of the terms within the current events documents;

generating a centroid feature vector for the subset of the news feature vectors, based on the sets of numerical values of the subset of the news feature vectors;

forming the augmented query feature vector, based on a comparison of the query feature vector and the centroid feature vector;

accessing target feature vectors associated with target documents, each the target feature vectors comprising sets of numerical values associated with terms in corresponding ones of the target documents;

computing first metrics of similarity between the augmented query feature vector and the target feature vectors, the first similarity metrics comprising at least one of distances or angles between the augmented query feature vector and the target feature vectors;

identifying search results based on the computed first similarity metrics, the search results comprising information associated with at least a portion of the target documents; and

enabling the user to perceive at least one of the identified search results.