Automatic tagging of content based on a corpus of previously tagged and untagged content

• US 20080228749A1
• Filed: 03/13/2007
• Published: 09/18/2008
• Est. Priority Date: 03/13/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Brown, Robert I.
• Granted Patent: US 8,103,646 B2
• Time in Patent Office: Days
• US Class Current: 707/5
• CPC Class Codes: G06F 16/685 using automatically derived transcript of audio data, e.g. lyrics speech recognition G10L15/00
• CPC Class Codes: G06F 16/78 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06F 16/7844 using original textual content or text extracted from visual content or transcript of audio data

Abstract

An automated mechanism of automatically tagging media files such as podcasts, blog entries, and videos, for example, with meaningful taxonomy tags. The mechanism provides active (or automated) assistance in assigning appropriate tags to a particular piece of content (or media). Included is a system for automatic tagging of audio streams on the Internet, whether from audio files, or from the audio tracks of audio/video files, using the folksonomy of the Internet. The audio streams may be provided by the media author. For example, the author can make a recording to be posted on a website, and use the system to automatically suggest (via prompted author interaction) folksonomically appropriate tags for the media recording. Alternatively, the system can be used in an automated fashion to develop and assign without any intervention by the author.

Claims

1. A computer-implemented method of managing information, comprising:

searching information sources for tag and content relationship data;

creating a tag model based on the relationship data; and

generating a tag for new content based on the tag model.

2. The method of claim 1, further comprising classifying text data from the information sources for creation of the tag model and generating the tag based on the tag model.

3. The method of claim 1, further comprising generating a ranked list of tags from which the tag is selected for the new content.

4. The method of claim 3, further comprising automatically selecting the tag from the ranked list and assigning the tag to the new content.

5. The method of claim 3, further comprising manually selecting the tag from the ranked list and assigning the tag to the new content.

6. The method of claim 1, further comprising classifying text data from the information sources based on a confidence threshold processed by a classifier.

7. The method of claim 1, further comprising:

accessing a speech model to transcribe audio and/or video content; and

training the speech model based on the audio and/or video content.

8. The method of claim 1, further comprising:

training the tag model based on tagged content and blog text;

classifying recognized tag and content relationship data based on the tag model; and

generating a list of ranked tags for the new content, the tag selected from the list of ranked tags.

9. The method of claim 1, further comprising;

training the tag model based on at least one of tagged content, transcribed untagged audio and/or video data, tagged audio and/or video data, or blog text; and

generating a list of ranked tags for the new content based on folksonomy, as defined by the tag model.

10. A computer-implemented system that facilitates information management, comprising:

a modeling component for generating a model of taxonomy for tags and an associated corpus of content; and

a tagging component for automatically generating tags for new content based on the model.

11. The system of claim 10, wherein the modeling component generates the model based on content that includes untagged audio data.

12. The system of claim 10, wherein the modeling component generates the model based on content that includes untagged blog posting data.

13. The system of claim 10, wherein the modeling component generates the model based on audio content associated with video data.

14. The system of claim 10, further comprising a tag classifier for utilizing model information from the model and generating the tags.

15. The system of claim 10, further comprising a model trainer for training at least one of a speech recognition model or a tag classification model.

16. The system of claim 10, wherein the tagging component generates the tags based on at least one of human-assigned tags, source data, or blog posting text.

17. The system of claim 10, further comprising a transcriber for automatically transcribing audio data obtained from a network site into text for classification.

18. The system of claim 10, further comprising a network crawler for automatically searching sites for the tags and the associated corpus of the content.

19. The system of claim 10, further comprising a machine learning and reasoning component that employs a probabilistic and/or statistical-based analysis to prognose or infer an action that is desired to be automatically performed.

20. A computer-implemented system, comprising:

computer-implemented means for searching information sources for tag and content relationship data;

computer-implemented means for creating a tag model based on the relationship data; and

computer-implemented means for generating a tag for new content based on the tag model.