Generating snippet modules on online social networks

• US 10,509,832 B2
• Filed: 07/13/2015
• Issued: 12/17/2019
• Est. Priority Date: 07/13/2015
• Status: Active Grant

• Current Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Original Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Inventors: Sauper, Christina Joan, Do, Quang Xuan, Bao, Ling, Eldar, Bosmat Mazal, Bellare, Kedar
• Primary Examiner(s): Harper, Eliyah S.
• Application Number:
  US14/797,819
• Publication Number:
  US 20170017721A1
• Time in Patent Office: 1,618 Days
• Field of Search: 707784
• CPC Class Codes: G06F 16/24575 using context
• CPC Class Codes: G06F 16/248 Presentation of query results
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

In one embodiment, a method includes accessing posts of an online social network, each post being associated with a first entity of the online social network, classifying, based on content and metadata associated with each post, one or more of the posts as being relevant to the first entity, extracting a set of one or more n-grams from the content of the posts classified as being relevant to the first entity, filtering the set of n-grams to remove one or more of the extracted n-grams from the set of n-grams, calculating a quality score for each n-gram in the filtered set of n-grams, generating a snippet-module including one or more of the extracted n-grams from the filtered set of n-grams having quality-scores greater than a threshold quality-score, and sending, to a client system of a first user of the online social network, the snippet-module for display to the user.

Claims

1. A method comprising, by one or more computing devices:

receiving, at the one or more computing devices from a client system associated with a first user of the online social network, a request to access a profile-page for a first entity of the online social network, the first entity being of a particular entity-type;

accessing, by a snippet-generation system of the one or more computing devices, a plurality of posts of the online social network, each post being associated with the first entity of the online social network, wherein each post comprises content of the post and metadata of the post, and wherein each post is linked to the profile-page of the first entity;

accessing, by the one or more computing devices, a set of pre-determined n-grams linked to the particular entity-type of the first entity, wherein the pre-determined n-grams are n-grams that have been determined to be associated with high-quality content about entities of the particular entity-type;

classifying, by a classifier process of the snippet-generation system, one or more of the posts as being relevant to the first entity, each post being classified based on a comparison of pre-determined n-grams and the content of the post and the metadata of the post;

generating, by a noun-phrase extractor process of the snippet-generation system, a set of n-grams by extracting one or more n-grams from the content of the one or more of the posts classified as being relevant to the first entity;

filtering, by a filter process of the snippet-generation system, the set of n-grams to remove one or more of the extracted n-grams from the set of n-grams;

calculating, by a score calculator process of the snippet-generation system, for each n-gram in the filtered set of n-grams, a quality-score for the n-gram;

generating, by the snippet-generation system, a snippet-module comprising one or more of the extracted n-grams from the filtered set of n-grams having quality-scores greater than a threshold quality-score, wherein the snippet-module is generated based on the particular entity-type of the first entity and references the posts that the extracted n-grams are associated with; and

sending, from the one or more client systems to the client system of the first user, in response to the request, instructions for presenting the profile-page of the first entity, wherein the profile-page comprises the snippet-module.

2. The method of claim 1, further comprising:

accessing a social graph comprising a plurality of nodes and a plurality of edges connecting the nodes, each of the edges between two of the nodes representing a single degree of separation between them, the nodes comprising;

a first node corresponding to the first entity; and

a plurality of second nodes corresponding to a plurality of second entities associated with the online social network, respectively.

3. The method of claim 1, wherein the plurality of posts comprise one or more of:

a publication by a second entity referencing the first entity in a profile page of the second entity on the online social network;

or

a publication by a second entity referencing the first entity in a profile page of the first entity on the online social network.

4. The method of claim 1, wherein the plurality of posts comprise one or more of:

a publication by a second entity referencing the first entity in a profile page of a third entity on the online social network.

5. The method of claim 1, wherein classifying each post as being relevant to the first entity comprises:

applying a classifier algorithm to each post to determine whether the post is relevant, wherein the classifier algorithm is trained to identify relevant posts.

6. The method of claim 1, wherein classifying each post as being relevant to the first entity based on the content of the post comprises:

parsing the content of the post to extract one or more n-grams from the content of the post;

comparing the extracted n-grams with a pre-determined cluster of n-grams linked to an entity-type of the first entity; and

classifying the post as being relevant to the first entity based at least on the comparison of the extracted n-grams with the pre-determined cluster of n-grams.

7. The method of claim 1, wherein classifying each post as being relevant to the first entity comprises:

calculating a relevance-score for the post based on the content of the post and the metadata of the post; and

classifying the post as being relevant to the first entity if the post has a relevance-score above a pre-determined threshold relevance-score.

8. The method of claim 1, wherein the extracted n-grams comprise one or more noun phrases.

9. The method of claim 8, wherein each of the noun phrases comprises:

at least one noun;

at least one noun and at least one adjective;

or

at least one noun and at least one conjunction.

10. The method of claim 1, wherein filtering the set of n-grams to remove one or more of the extracted n-grams from the set of n-grams comprises removing one or more of the following n-grams from the set of n-grams:

one or more noun phrases associated with one or more second entities;

one or more redundant noun phrases;

one or more non-informative noun phrases;

one or more brand names;

one or more noun phrases appearing only in non-public posts of the plurality of posts;

one or more noun phrases that are not associated with a dictionary of noun phrases;

or any combinations thereof.

11. The method of claim 10, wherein a non-informative noun phrase comprises a noun associated with profanity.

12. The method of claim 10, wherein the redundant noun phrases comprise one or more overlapped noun phrases having Jaccard'"'"'s similarity coefficient values above a pre-determined threshold coefficient value.

13. The method of claim 1, wherein filtering the set of n-grams to remove one or more of the extracted n-grams from the set of n-grams comprises removing one or more of the following n-grams from the set of n-grams:

one or more noun phrases that appear only in social-networking data associated with the first entity;

one or more noun phrases that appear in less than a threshold number of posts of the plurality of posts;

one or more noun phrases that are authored by not more than a pre-determined number of users of the online social network;

or

any combinations thereof.

14. The method of claim 1, wherein the quality-score for each n-gram in the filtered set of n-grams is calculated based at least on a term frequency-inverse document frequency (TF-IDF) algorithm, the TF-IDF algorithm taking as an input a number of times the n-gram appears in first social-networking data associated with the first entity and a number of times the n-gram appears in second social-networking data associated with a plurality of entities of the online social network associated with an entity-type of the first entity.

15. The method of claim 1, wherein the quality-score for each n-gram in the filtered set of n-grams is calculated based at least on a measure of privacy of the n-gram.

16. The method of claim 1, wherein the quality-score for each n-gram in the filtered set of n-grams is calculated based at least on a number of uni-grams in the n-gram.

17. The method of claim 1, wherein the quality-score for each n-gram in the filtered set of n-grams is calculated based at least on a comparison of an adjective n-gram in the n-gram with a pre-determined cluster of adjective n-grams linked to an entity-type of the first entity.

20. The method of claim 1, wherein the profile-page of the first entity is a webpage accessed by a browser client on the client system.

21. The method of claim 1, wherein the profile-page of the first entity is a user interface of a native application installed on the client system.

18. One or more computer-readable non-transitory storage media embodying software that is operable when executed to:

receive, at the one or more computing devices from a client system associated with a first user of the online social network, a request to access a profile-page for a first entity of the online social network, the first entity being of a particular entity-type;

access, by a snippet-generation system of the one or more computing devices, a plurality of posts of the online social network, each post being associated with the first entity of the online social network, wherein each post comprises content of the post and metadata of the post, and wherein each post is linked to the profile-page of the first entity;

access, by the one or more computing devices, a set of pre-determined n-grams linked to the particular entity-type of the first entity, wherein the pre-determined n-grams are n-grams that have been determined to be associated with high-quality content about entities of the particular entity-type;

classify, by a classifier process of the snippet-generation system, one or more of the posts as being relevant to the first entity, each post being classified based on a comparison of pre-determined n-grams and the content of the post and the metadata of the post;

generate, by a noun-phrase extractor process of the snippet-generation system, a set of n-grams by extracting one or more n-grams from the content of the one or more of the posts classified as being relevant to the first entity;

filter, by a filter process of the snippet-generation system, the set of n-grams to remove one or more of the extracted n-grams from the set of n-grams;

calculate, by a score calculator process of the snippet-generation system, for each n-gram in the filtered set of n-grams, a quality-score for the n-gram;

generate, by the snippet-generation system, a snippet-module comprising one or more of the extracted n-grams from the filtered set of n-grams having quality-scores greater than a threshold quality-score, wherein the snippet-module is generated based on the particular entity-type of the first entity and references the posts that the extracted n-grams are associated with; and

send, from the one or more client systems to the client system of the first user, in response to the request, instructions for presenting the profile-page of the first entity, wherein the profile-page comprises the snippet-module.

19. A system comprising:

one or more processors; and

a memory coupled to the processors comprising instructions executable by the processors, the processors operable when executing the instructions to;

receive, at the one or more computing devices from a client system associated with a first user of the online social network, a request to access a profile-page for a first entity of the online social network, the first entity being of a particular entity-type;

access, by a snippet-generation system of the one or more computing devices, a plurality of posts of the online social network, each post being associated with the first entity of the online social network, wherein each post comprises content of the post and metadata of the post, and wherein each post is linked to the profile-page of the first entity;

access, by the one or more computing devices, a set of pre-determined n-grams linked to the particular entity-type of the first entity, wherein the pre-determined n-grams are n-grams that have been determined to be associated with high-quality content about entities of the particular entity-type;

classify, by a classifier process of the snippet-generation system, one or more of the posts as being relevant to the first entity, each post being classified based on a comparison of pre-determined n-grams and the content of the post and the metadata of the post;

generate, by a noun-phrase extractor process of the snippet-generation system, a set of n-grams by extracting one or more n-grams from the content of the one or more of the posts classified as being relevant to the first entity;

filter, by a filter process of the snippet-generation system, the set of n-grams to remove one or more of the extracted n-grams from the set of n-grams;

calculate, by a score calculator process of the snippet-generation system, for each n-gram in the filtered set of n-grams, a quality-score for the n-gram;

generate, by the snippet-generation system, a snippet-module comprising one or more of the extracted n-grams from the filtered set of n-grams having quality-scores greater than a threshold quality-score, wherein the snippet-module is generated based on the particular entity-type of the first entity and references the posts that the extracted n-grams are associated with; and

send, from the one or more client systems to the client system of the first user, in response to the request, instructions for presenting the profile-page of the first entity, wherein the profile-page comprises the snippet-module.

22. The system of claim 19, wherein the processors are further operable when executing the instructions to:

access a social graph comprising a plurality of nodes and a plurality of edges connecting the nodes, each of the edges between two of the nodes representing a single degree of separation between them, the nodes comprising;

a first node corresponding to the first entity; and

a plurality of second nodes corresponding to a plurality of second entities associated with the online social network, respectively.

23. The system of claim 19, wherein the plurality of posts comprise one or more of:

a publication by a second entity referencing the first entity in a profile page of the second entity on the online social network;

or

a publication by a second entity referencing the first entity in a profile page of the first entity on the online social network.

24. The system of claim 19, wherein the plurality of posts comprises one or more of:

a publication by a second entity referencing the first entity in a profile page of a third entity on the online social network.

25. The system of claim 19, wherein the instruction to classify each post as being relevant to the first entity comprises instructions to:

apply a classifier algorithm to each post to determine whether the post is relevant, wherein the classifier algorithm is trained to identify relevant posts.

26. The system of claim 19, wherein the instructions to classify each post as being relevant to the first entity comprises instructions to:

parse the content of the post to extract one or more n-grams from the content of the post;

compare the extracted n-grams with a pre-determined cluster of n-grams linked to an entity-type of the first entity; and

classify the post as being relevant to the first entity based at least on the comparison of the extracted n-grams with the pre-determined cluster of n-grams.

27. The system of claim 19, wherein the instructions to classify each post as being relevant to the first entity comprises instructions to:

calculate a relevance-score for the post based on the content of the post and the metadata of the post; and

classify the post as being relevant to the first entity if the post has a relevance-score above a pre-determined threshold relevance-score.

28. The system of claim 19, wherein the extracted n-grams comprise one or more noun phrases.

29. The system of claim 28, wherein each of the noun phrases comprises:

at least one noun;

at least one noun and at least one adjective;

or

at least one noun and at least one conjunction.

30. The system of claim 19, wherein the instructions to filter the set of n-grams to remove one or more of the extracted n-grams from the set of n-grams comprises instructions to remove one or more of the following n-grams from the set of n-grams:

one or more noun phrases associated with one or more second entities;

one or more redundant noun phrases;

one or more non-informative noun phrases;

one or more brand names;

one or more noun phrases appearing only in non-public posts of the plurality of posts;

one or more noun phrases that are not associated with a dictionary of noun phrases;

or any combinations thereof.

31. The system of claim 30, wherein a non-informative noun phrase comprises a noun associated with profanity.

32. The system of claim 30, wherein the redundant noun phrases comprise one or more overlapped noun phrases having Jaccard'"'"'s similarity coefficient values above a pre-determined threshold coefficient value.

33. The system of claim 19, wherein the instructions to filter the set of n-grams to remove one or more of the extracted n-grams from the set of n-grams comprises instructions to remove one or more of the following n-grams from the set of n-grams:

one or more noun phrases that appear only in social-networking data associated with the first entity;

one or more noun phrases that appear in less than a threshold number of posts of the plurality of posts;

one or more noun phrases that are authored by not more than a pre-determined number of users of the online social network;

or

any combinations thereof.

34. The system of claim 19, wherein the quality-score for each n-gram in the filtered set of n-grams is calculated based at least on a term frequency-inverse document frequency (TF-IDF) algorithm, the TF-IDF algorithm taking as an input a number of times the n-gram appears in first social-networking data associated with the first entity and a number of times the n-gram appears in second social-networking data associated with a plurality of entities of the online social network associated with an entity-type of the first entity.

35. The system of claim 19, wherein the quality-score for each n-gram in the filtered set of n-grams is calculated based at least on a measure of privacy of the n-gram.

36. The system of claim 19, wherein the quality-score for each n-gram in the filtered set of n-grams is calculated based at least on a number of uni-grams in the n-gram.

37. The system of claim 19, wherein the quality-score for each n-gram in the filtered set of n-grams is calculated based at least on a comparison of an adjective n-gram in the n-gram with a pre-determined cluster of adjective n-grams linked to an entity-type of the first entity.

38. The system of claim 19, wherein the profile-page of the first entity is a webpage accessed by a browser client on the client system.

39. The system of claim 19, wherein the profile-page of the first entity is a user interface of a native application installed on the client system.