Methods, systems and techniques for ranking personalized and generic search query suggestions

1. A method, implemented on at least one computing device each of which has at least one processor, storage, and a communication platform connected to a network to provide query suggestions, the method comprising:

receiving, in a person session involving a person interacting with a search application, a portion of a search query from the person;

sending the portion of the search query to a non-personalized database, to obtain a set of non-personalized query suggestions based on the portion of the search query without considering a previous query within the person session;

looking up, completely in parallel to obtaining the set of non-personalized query suggestions, based on the portion of the search query, a personalized cache associated with a person corpus for a set of personalized query suggestions based on the portion of the search query, wherein the person corpus is derived based on a person-centric space including data cross linked from a private space associated with the person, a semi-private space, and a public space based on cross-linking keys identified from the private space;

in response to a failure of obtaining the set of personalized query suggestions from the personalized cache, sending, asynchronously with the looking up, a request to the person corpus for personalized data to be written into the personalized cache;

in response to receipt of an additional portion of the search query from the person, obtaining the set of personalized query suggestions from the personalized cache;

ranking, based on a usage of a first application associated with the person corpus, query suggestions from the first set of personalized query suggestions and the second set of non-personalized query suggestions to generate a ranked list of query suggestions, wherein a query suggestion is ranked higher if a first frequency that the person accesses the first application exceeds a threshold within a period of time and the first frequency is higher than a second frequency that the person accesses a second application associated with the person corpus; and

presenting at least some of the ranked query suggestions to the person, so as to facilitate the person to submit a complete search query.

2. The method of claim 1 , wherein the private space includes:

private electronic mails;

a private calendar;

a private contact list;

private messages;

private bookmarks;

private visual information;

private voice information; and

private log information.

3. The method of claim 1 , wherein the non-personalized database includes:

queries mined from query logs with respect to a general population of persons without considering the person's current search behavior.

4. The method of claim 1 , further comprising:

mapping attributes of each query suggestion of the first set of personalized query suggestions and the set of non-personalized query suggestions to a feature space, wherein the operation of ranking comprises:

normalizing each of the mapped query suggestions in the feature space to generate a normalized feature set;

estimating a score for each query suggestion based on the normalized feature set; and

ranking each query suggestion from the set of personalized query suggestions and the set of non-personalized query suggestions based on the score of the query suggestion.

5. The method of claim 1 , wherein the portion includes a prefix of a search query.

6. The method of claim 1 , further comprising:

presenting the ranked set of personalized query suggestions prior to the ranked set of non-personalized query suggestions.

7. The method of claim 1 , further comprising:

blending the set of personalized query suggestions and the set of non-personalized query suggestions to generate the ranked list of query suggestions; and

presenting at least one suggestion out of the ranked list of suggestions with the score.

8. The method of claim 1 , wherein each query suggestion of the set of personalized query suggestions and the set of non-personalized query suggestions is mapped to the feature space based on at least one attribute associated with the query suggestion, the attribute indicating at least one statistic related to the query suggestion.

9. The method of claim 1 , wherein the model to extract the query suggestion from the data source is selected based on a type of content included in the data source.

10. The method of claim 1 , wherein deriving the person corpus further comprises:

determining a domain of the data; and

identifying the cross-linking keys from the data based on the domain.

11. The method of claim 1 , wherein the model is trained to extract the query suggestion based on a type of information.

12. A system having at least one processor, storage, and a communication platform, to provide query suggestions, the system comprising:

a request processing unit implemented using the at least one processor and configured to:

receive, in a person session involving a person interacting with a search application, a portion of a search query from a person;

a suggestion search unit implemented using the at least one processor and configured to:

send the portion of the search query to a non-personalized database, to obtain a set of non-personalized query suggestions based on the portion of the search query without considering a previous query within the person session,

look up, completely in parallel to obtaining the set of non-personalized query suggestions, based on the portion of the search query, a personalized cache associated with a person corpus for a set of personalized query suggestions based on the portion of the search query, wherein the person corpus is derived based on a person-centric space including data cross linked from a private space associated with the person, a semi-private space, and a public space based on cross-linking keys identified from the private space,

in response to a failure of obtaining the set of personalized query suggestions from the personalized cache, send, asynchronously with the looking up, a request to the person corpus for personalized data to be written into the personalized cache, and

in response to receipt of an additional portion of the search query from the person, obtain the set of personalized query suggestions from the personalized cache; and

a blended query suggestion generation module implemented using the at least one processor and configured to:

rank, based on a usage of a first application associated with the person corpus, query suggestions from the first set of personalized query suggestions and the second set of non-personalized query suggestions to generate a ranked list of query suggestions, wherein a query suggestion is ranked higher if a first frequency that the person accesses the first application exceeds a threshold within a period of time and the first frequency is higher than a second frequency that the person accesses a second application associated with the person corpus, and

provide at least some of the ranked query suggestions to the person, so as to facilitate the person to submit a complete search query.

13. The system of claim 12 , wherein the private space includes:

private electronic mails;

a private calendar;

a private contact list;

private messages;

private bookmarks;

private visual information;

private voice information; and

private log information.

14. The system of claim 12 , wherein the non-personalized database includes:

queries mined from query logs with respect to a general population of persons without considering the person's current search behavior.

15. The system of claim 12 , wherein the blended query suggestion generation module is configured to:

map attributes of each query suggestion of the set of personalized query suggestions and the set of non-personalized query suggestions to a feature space;

normalize each of the mapped query suggestions in the feature space to generate a normalized feature set;

estimate a score for each query suggestion based on the normalized feature set; and

rank each query suggestion from the set of personalized query suggestions and the set of non-personalized query suggestions based on the score of the query suggestion.

16. The system of claim 12 , wherein the portion includes a prefix of a search query.

17. The system of claim 12 , wherein the blended query suggestion generation module is further configured to present the ranked set of query suggestions prior to the ranked set of query suggestions.

18. A non-transitory computer-readable storage medium having information recorded thereon for providing query suggestions, wherein the information, when read by a machine, causes the machine to perform operations comprising:

receiving, in a person session involving a person interacting with a search application, a portion of a search query from a person;

sending the portion of the search query to a non-personalized database, to obtain a set of non-personalized query suggestions based on the portion of the search query without considering a previous query within the person session;

looking up, completely in parallel to obtaining the set of non-personalized query suggestions, based on the portion of the search query, a personalized cache associated with a person corpus for a set of personalized query suggestions based on the portion of the search query, wherein the person corpus is derived based on a person-centric space including data cross linked from a private space associated with the person, a semi-private space, and a public space based on cross-linking keys identified from the private space;

in response to a failure of obtaining the set of personalized query suggestions from the personalized cache, sending, asynchronously with the looking up, a request to the person corpus for personalized data to be written into the personalized cache;

in response to receipt of an additional portion of the search query from the person, obtaining the set of personalized query suggestions from the personalized cache;

ranking, based on a usage of a first application associated with the person corpus, query suggestions from the first set of personalized query suggestions and the second set of non-personalized query suggestions to generate a ranked list of query suggestions, wherein a query suggestion is ranked higher if a first frequency that the person accesses the first application exceeds a threshold within a period of time and the first frequency is higher than a second frequency that the person accesses a second application associated with the person corpus; and

presenting at least some of the ranked query suggestions to the person, so as to facilitate the person to submit a complete search query.

19. The medium of claim 18 , wherein the private space includes:

private electronic mails;

a private calendar;

a private contact list;

private messages;

private bookmarks;

private visual information;

private voice information; and

private log information.

20. The medium of claim 18 , wherein the non-personalized database includes:

queries mined from query logs with respect to a general population of persons without considering the person's current search behavior.

21. The medium of claim 18 , wherein the information, when read by the machine, causes the machine to further perform mapping attributes of each query suggestion of the set of personalized query suggestions and the set of non-personalized query suggestions to a feature space, and wherein the operation of ranking comprises:

normalizing each of the mapped query suggestions in the feature space to generate a normalized feature set;

estimating a score for each query suggestion based on the normalized feature set; and

ranking each query suggestion from the set of query suggestions and the set of query suggestions based on the score of the query suggestion.

22. The medium of claim 18 , wherein the portion includes a prefix of a search query.