Search queries with shifting intent
Techniques and systems are disclosed for returning temporally-aware results from an Internet-based search query. To determine if a query is temporally-based one or more query features are collected and input into a trained classifier, yielding a temporal classification for the query. Further, if a query is classified as temporal, the query results are shifted by determining an alternate set of results for the query, and returning one or more alternate results to one or more users. Based on user interactions with the one or more alternate results, the classifier can be updated, for example, by changing the query to a non-temporal query if the user interactions identify it as such.
1. A computer-implemented method for returning temporally aware results for a query, comprising:
identifying a first set of results for a query, the first set of results non-temporally based;
classifying whether the query is temporally based by inputting one or more features of the query into a trained classifier to assign a classification to the query;
if the query is classified as temporally based, determining an alternate set of results for the query different than the first set of results; and
updating the classification based upon user interaction with the alternate set of results, comprising:
representing the one or more features of the query as a vector of coordinates in feature space, the vector of coordinates assigned a temporally based label if the query is classified as temporally based; and
if the user interaction with the alternate set of results indicates the query is non-temporally based and the vector of coordinates is assigned the temporally based label, then:
modifying the temporally based label assigned to the vector of coordinates to a non-temporally based label; and
updating a margin separating a temporally-based query region and a non-temporally based query region in the feature space, at least some of at least one of the identifying, the classifying, the determining, and the updating the classification implemented at least in part via a processor.
2. The method of claim 1 , comprising:
assigning the temporally based label to the vector of coordinates if the vector of coordinates is within the temporally-based query region of the feature space.
3. The method of claim 1 , comprising:
assigning the non-temporally based label to the vector of coordinates if the vector of coordinates is within the non-temporally based query region of the feature space.
4. The method of claim 1 , comprising:
collecting one or more features of the query comprising at least one of:
determining how a user interacts with the first set of results returned for the query;
determining characteristics of the query; and
determining statistical-based characteristics of one or more of:
burstiness of user interactions with the first set of results returned for the query;
geographical location related to the query;
age of a document resulting from the query;
age of a selected document in the first set of results;
a desired number of recent queries that are similar to the query; and
a desired number of recent news queries that are similar to the query.
5. The method of claim 1 , the determining an alternate set of results comprising:
using a search engine to identify the alternate set of results comprising different characteristics than characteristics of the first set of results.
6. The method of claim 1 , the updating the classification comprising:
comparing the user interaction with the alternate set of results to initial user interactions with the first set of results.
7. The method of claim 6 , the updating the classification comprising:
if the user interaction with the alternate set of results deviates from the initial user interactions with the first set of results below a threshold margin, then classifying the query as non-temporally based.
8. The method of claim 1 , the updating the classification comprising:
if characteristics of the alternate set of results deviates from characteristics of the first set of results below a threshold margin, then classifying the query as non-temporally based.
9. A computer-implemented system for returning temporally aware results for a query, comprising:
an alternate results determination component configured to determine an alternate set of results for a query different than a first set of results for the query if the query is classified as temporally based;
a feature space function component configured to:
represent one or more features of the query as a vector of coordinates in feature space, the vector of coordinates assigned a temporally based label if the query is classified as temporally based; and
if user interaction with the alternate set of results indicates the query is non-temporally based and the vector of coordinates is assigned the temporally based label, then:
modify the temporally based label assigned to the vector of coordinates to a non-temporally based label; and
update a margin separating a temporally-based query region and a non-temporally based query region in the feature space, at least some of at least one of the alternate results determination component and the feature space function component implemented at least in part via a processor.
10. The system of claim 9 , the first set of results non-temporally based.
11. The system of claim 9 , the feature space function component configured to:
assign the temporally based label to the vector of coordinates if the vector of coordinates is within the temporally-based query region of the feature space.
12. The system of claim 9 , the feature space function component configured to:
assign the non-temporally based label to the vector of coordinates if the vector of coordinates is within the non-temporally based query region of the feature space.
13. A computer readable storage medium comprising instructions that when executed perform a method for returning temporally aware results for a query, comprising:
identifying a first set of results for a query, the first set of results non-temporally based;
classifying whether the query is temporally based by inputting one or more features of the query into a trained classifier to assign a classification to the query;
if the query is classified as temporally based, determining an alternate set of results for the query different than the first set of results; and
updating the classification based upon user interaction with the alternate set of results, comprising:
representing the one or more features of the query as a vector of coordinates in feature space, the vector of coordinates assigned a temporally based label if the query is classified as temporally based; and
if the user interaction with the alternate set of results indicates the query is non-temporally based and the vector of coordinates is assigned the temporally based label, then:
modifying the temporally based label assigned to the vector of coordinates to a non-temporally based label; and
updating a margin separating a temporally-based query region and
a non-temporally based query region in the feature space.
14. The computer readable storage medium of claim 13 , the method comprising:
collecting one or more features of the query comprising at least one of:
determining how a user interacts with the first set of results returned for the query;
determining characteristics of the query; and
determining statistical-based characteristics of one or more of:
burstiness of user interactions with the first set of results returned for the query;
geographical location related to the query;
age of a document resulting from the query;
age of a selected document in the first set of results;
a desired number of recent queries that are similar to the query; and
a desired number of recent news queries that are similar to the query.
15. The computer readable storage medium of claim 13 , the method comprising:
returning a first alternate result comprised within the alternate set of results to a first user and a second alternate result comprised within the alternate set of results to a second user.
16. The computer readable storage medium of claim 13 , one or more characteristics associated with the alternate set of results different than one or more characteristics associated with the first set of results.
17. The computer readable storage medium of claim 13 , the method comprising:
re-classifying the query as non-temporally based if one or more characteristics associated with the alternate set of results deviate from one or more characteristics associated with the first set of results below a threshold margin.
18. The computer readable storage medium of claim 13 , the method comprising:
using a search engine to identify the alternate set of results.
19. The computer readable storage medium of claim 13 , the method comprising:
returning one or more results based upon a comparison of one or more characteristics associated with the alternate set of results to one or more characteristics associated with the first set of results.
20. The computer readable storage medium of claim 13 , one or more characteristics associated with the alternate set of results comprising an interaction with the alternate set of results, one or more characteristics associated with the first set of results comprising an initial interaction with the first set of results.