Non-text content item search

• US 8,275,771 B1
• Filed: 06/07/2010
• Issued: 09/25/2012
• Est. Priority Date: 02/26/2010
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Malpani, Radhika, Preetham, Arcot J., Mate, Omkar
• Primary Examiner(s): Vy, Hung T
• Application Number:
  US12/795,239
• Time in Patent Office: 841 Days
• Field of Search: 707/736, 707/737, 707/738, 707/749, 707/776
• US Class Current: 707/737
• CPC Class Codes: G06F 16/95 Retrieval from the web
• CPC Class Codes: G06F 16/954 Navigation, e.g. using categorised browsing

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for selecting labels for a non-text content item. In one aspect, a method receives a set of initial labels for a non-text content item, wherein the set of initial labels specifies text that has been identified as descriptive of the non-text content item and a web page to which the text corresponds. Initial labels corresponding to sets of matching web pages are grouped into separate initial label groups that correspond to each set of matching web pages. Sets of matching labels are grouped into other separate initial label groups that correspond to the sets of matching labels. One or more words that are included in at least a threshold number of the separate label groups are selected as final labels for the non-text content item.

Claims

1. A system, comprising:

a data store storing label data that specifies a set of initial labels for a non-text content item and a resource identifier that specifies, for each initial label, a web page to which the initial label is associated, wherein the non-text content item is associated with each of a plurality of web pages, and wherein each initial label includes one or more words; and

one or more computers coupled to the data store, the one or more computers configured to;

generate initial label groups for sets of matching web pages, each set of matching web pages including two or more matching web pages;

group, for each set of matching web pages, initial labels that are associated with the set of matching web pages into a separate initial label group that corresponds to the set of matching web pages;

generate initial label groups for sets of matching labels, each set of matching labels including two or more initial labels;

group, for each set of matching labels, initial labels that are associated with the set of matching labels into a separate initial label group that corresponds to the set of matching labels; and

select, as final labels for the non-text content item, n-grams of one or more words that are included in at least a threshold number of the separate initial label groups.

2. The system of claim 1, wherein the non-text content item comprises content that is provided in a non-text format.

3. The system of claim 1, wherein the non-text content item comprises at least one of audio, video, and multimedia content.

4. The system of claim 1, wherein the one or more computers are further configured to select, as a set of matching web pages, web pages corresponding to a same domain.

5. The system of claim 1, wherein the one or more computers are further configured to select, as a set of matching web pages, web pages that are provided for a same publisher.

6. The system of claim 1, wherein the one or more computers are further configured to select, as a set of matching labels, initial labels having a label similarity score that exceeds a label similarity threshold.

7. The system of claim 1, wherein the one or more computers are further configured to select, as a set of matching labels, initial labels that each include at least one same word.

8. The system of claim 1, wherein the one or more computers are further configured to select, as a set of matching labels, initial labels that have at least a specified threshold cosine similarity measure.

9. The system of claim 1, wherein the one or more computers are further configured to receive a search query, determine that a final label matches the search query, and adjust, using a boost factor, a baseline result score for content corresponding to the final label that matches the search query.

10. The system of claim 9, further comprising a user device, wherein the one or more computers are further configured to select a presentation order for search results using the adjusted result scores and provide data to the user device that causes presentation of the search results by the user device and according to the presentation order.

11. The system of claim 1, wherein the set of initial labels comprises text that is provided for presentation within a threshold number of pixels of the non-text content item.

12. The system of claim 1, wherein the set of initial labels comprises metadata for the web page with which the initial label is associated.

13. A non-transitory computer storage medium encoded with a computer program, the program comprising instructions that when executed by data processing apparatus cause the data processing apparatus to perform operations comprising:

receiving label data that specifies a set of initial labels for a non-text content item and a resource identifier for each initial label, wherein each initial label includes one or more words;

selecting one or more sets of matching web pages from a plurality of web pages, wherein each set of matching web pages includes two or more matching web pages;

grouping, for each set of matching web pages, initial labels that are associated with the set of matching web pages into a separate initial label group that corresponds to the set of matching web pages;

selecting one or more sets of matching labels, wherein each set of matching labels includes two or more initial labels;

grouping each set of matching labels into a separate initial label group that corresponds to the set of matching labels; and

selecting, as a final label for the non-text content item, an n-gram of one or more words that are included in at least a threshold number of separate initial label groups.

14. The non-transitory computer storage medium of claim 13, wherein receiving label data specifying a set of initial labels for a non-text content item comprises receiving a set of initial labels for content that is provided in a non-text format.

15. The non-transitory computer storage medium of claim 13, wherein the program further comprises instructions that when executed by the data processing apparatus cause the data processing apparatus to perform operations comprising selecting, as a set of matching web pages, web pages that are provided for a same publisher.

16. The non-transitory computer storage medium of claim 13, wherein the program further comprises instructions that when executed by the data processing apparatus cause the data processing apparatus to perform operations comprising selecting, as a set of matching labels, initial labels having a label similarity score that exceeds a label similarity threshold.

17. A method performed by data processing apparatus, the method comprising:

selecting a non-text content item that is associated with each of a plurality of web pages;

receiving label data that includes a set of initial labels for the non-text content item and a resource identifier for each initial label, wherein each initial label includes one or more words;

selecting one or more sets of matching web pages from the plurality of web pages, wherein each set of matching web pages includes two or more matching web pages;

grouping, for each set of matching web pages, initial labels that are associated with the set of matching web pages into a separate initial label group that corresponds to the set of matching web pages;

selecting one or more sets of matching labels, wherein each set of matching labels includes two or more initial labels;

grouping each set of matching labels into a separate initial label group that corresponds to the set of matching labels; and

selecting, as a final label for the non-text content item, an n-gram of one or more words that are included in at least a threshold number of separate initial label groups.

18. The method of claim 17, further comprising:

receiving a search query;

determining that the search query matches the final label for the non-text content item; and

in response to the determination that the search query matches the final label, increasing a result score for the non-text content item, wherein the result score is a score with which a presentation order for non-text content items is selected.

19. The method of claim 18, further comprising:

selecting, based on the result scores, a presentation position for the non-text content item; and

providing data that cause presentation of the non-text content item at a presentation position according to the presentation order.

20. The method of claim 18, wherein determining that the search query matches the final label comprises determining that at least one same term is included in the search query and the final label.

21. The method of claim 18, wherein increasing the result score for the non-text content item comprises multiplying a baseline result score for the non-text content item by a boost factor having a value greater than 1.0.

22. The method of claim 21, further comprising:

computing a match score representing a measure of match between the search query and the final label; and

selecting the boost factor based on a value of the match score.

23. The method of claim 22, wherein selecting the boost factor comprises selecting a boost factor that is proportional to the match score.

24. The method of claim 17, wherein selecting a non-text content item that is associated with each of a plurality of web pages comprises selecting at least one of image content, audio content, video content, and multimedia content that is associated with each of a plurality of web pages.

25. The method of claim 17, wherein receiving label data that includes a set of initial labels comprises receiving text that is provided for presentation within a threshold number of pixels of the non-text content item.

26. The method of claim 17, wherein receiving label data that includes a set of initial labels comprises receiving metadata for the web page with which the non-text content item is associated.

27. The method of claim 17, wherein selecting sets of matching web pages comprises selecting web pages corresponding to a same domain.

28. The method of claim 17, wherein selecting sets of matching web pages comprises selecting web pages that are provided for a same publisher.

29. The method of claim 17, wherein selecting one or more sets of matching labels comprises selecting two or more labels that include at least one same word.

30. The method of claim 17, wherein selecting one or more sets of matching labels comprises selecting two or more labels that each include a word that corresponds to a same concept.

31. The method of claim 17, wherein selecting one or more sets of matching labels comprises selecting labels that have at least a threshold cosine similarity measure.

32. The method of claim 17, further comprising grouping at least one unique query into a separate initial label group, a unique query being a query that has not been grouped into one of an initial label group corresponding to a set of matching web pages and an initial label group corresponding to a set of matching labels.