MAILBOX SEARCH ENGINE USING QUERY MULTI-MODAL EXPANSION AND COMMUNITY-BASED SMOOTHING

• US 20140280207A1
• Filed: 03/15/2013
• Published: 09/18/2014
• Est. Priority Date: 03/15/2013
• Status: Active Grant

• Current Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Renders, Jean-Michel, Mantrach, Amin
• Granted Patent: US 9,280,587 B2
• Time in Patent Office: Days
• US Class Current: 707/748
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/3338 Query expansion
• CPC Class Codes: G06Q 10/00 Administration; Management
• CPC Class Codes: G06Q 10/107 Computer-aided management of electronic mailing [e-mailing]

Abstract

A retrieval method on a database of documents including text and names of participants associated with the documents includes: receiving a text query facet of keywords and a persons query facet of participant names; computing an enriched text query as an aggregation of the text query facet, a monomodal expansion of the text query facet based on the keywords, a cross-modal expansion of the text query facet based on the participant names, and a topic expansion of the text query facet based on a topic model associating words and topics; computing an enriched persons query as an aggregation of the persons query facet, a monomodal expansion of the persons query facet based on the participant names, a cross-modal expansion of the persons query facet based on the keywords, and a community expansion of the persons query facet based on a community model associating persons and communities.

Claims

1. A non-transitory computer-readable storage medium storing instructions executable by a computer to perform a retrieval method on a database of documents including text and names of participants associated with the documents by operations including:

receiving a multi-faceted retrieval query having a text query facet comprising one or more keywords and a persons query facet comprising one or more participant names;

computing an enriched text query as an aggregation of the text query facet, a monomodal expansion of the text query facet based on the one or more keywords, a cross-modal expansion of the text query facet based on the one or more participant names, and a topic expansion of the text query facet based on a topic model associating words and topics;

computing an enriched persons query as an aggregation of the persons query facet, a mono-modal expansion of the persons query facet based on the one or more participant names, a cross-modal expansion of the persons query facet based on the one or more keywords, and a community expansion of the persons query facet based on a community model associating persons and communities; and

performing ranking including at least one of;

(1) generating a ranking of documents by sorting similarities between the enriched text query and documents of the database, and

(2) generating a ranking of persons by sorting the enriched persons query.

2. The non-transitory computer-readable storage medium of claim 1, wherein the cross-modal expansion of the text query facet comprises text of person profiles of the named participants.

3. The non-transitory computer-readable storage medium of claim 1, wherein the cross-modal expansion of the text query facet is proportional to X_w^TX_uu_q where u_q is a vector representing the persons query facet, X_w is a term-document matrix associating word with documents of the database and X_u is a participant-document matrix associating participants with documents of the database.

4. The non-transitory computer-readable storage medium of claim 1, wherein the cross-modal expansion of the persons query facet comprises persons of social profiles comprising distribution of persons associated with the keywords.

5. The non-transitory computer-readable storage medium of claim 1, wherein the cross-modal expansion of the persons query facet is proportional to X_u^TX_ww_q where w_q is a vector representing the text query facet, X_w is a term-document matrix associating word with documents of the database and X_u is a participant-document matrix associating participants with documents of the database.

6. The non-transitory computer-readable storage medium of claim 1, wherein the performing of query enrichment further includes:

constructing a joint topic-community model defining hypertopics and associating topics with the hypertopics and communities with the hypertopics;

wherein the topic expansion uses the joint topic-community model as the topic model associating words and topics and the community expansion uses the joint topic-community model as the community model associating persons and communities.

7. The non-transitory computer-readable storage medium of claim 6, wherein the constructing of the joint topic-community model is performed offline before the performing of the query enrichment and the performing of the ranking.

8. The non-transitory computer-readable storage medium of claim 1, wherein the database of documents comprises an electronic mail (email) mailbox, the receiving comprises receiving the multi-faceted retrieval query from a computer running an email client, and the retrieval method further includes transmitting at least one of a ranking of documents and a ranking of persons to the computer running the email client.

9. An apparatus comprising:

a computer; and

a non-transitory computer-readable storage medium as set forth in claim 1 storing instructions executable by the computer to perform a retrieval method on a database of documents including text and names of participants associated with the documents.

10. A retrieval method operating on a database of documents including text and names of participants associated with the documents, the retrieval method comprising:

receiving a retrieval query having at least one of a text query facet comprising one or more keywords and a persons query facet comprising one or more participant names;

performing query enrichment including at least one of;

(1) computing an enriched text query as an aggregation of the text query facet, a monomodal expansion of the text query facet based on the one or more keywords, a cross-modal expansion of the text query facet based on the one or more participant names, and a topic expansion of the text query facet based on a topic model associating words and topics, and

(2) computing an enriched persons query as an aggregation of the persons query facet, a monomodal expansion of the persons query facet based on the one or more participant names, a cross-modal expansion of the persons query facet based on the one or more keywords, and a community expansion of the persons query facet based on a community model associating persons and communities; and

performing ranking including at least one of;

(1) generating a ranking of documents by sorting similarities between the enriched text query and documents of the database, and

(2) generating a ranking of persons by sorting the enriched persons query;

wherein the retrieval method is performed by a computer.

11. The retrieval method of claim 10 wherein:

the retrieval query includes both a text query facet and a persons query facet, and

the performing of query enrichment includes both computing an enriched text query and computing an enriched persons query.

12. The retrieval method of claim 10, wherein the cross-modal expansion of the text query facet comprises text of person profiles of the named participants.

13. The retrieval method of claim 10, wherein the cross-modal expansion of the text query facet is proportional to X_w^TX_uu_q where u_q is a vector representing the persons query facet, X_w is a term-document matrix associating word with documents of the database and X_u is a participant-document matrix associating participants with documents of the database.

14. The retrieval method of claim 10, wherein the cross-modal expansion of the persons query facet comprises persons of social profiles comprising distribution of persons associated with the keywords.

15. The retrieval method of claim 10, wherein the cross-modal expansion of the persons query facet is proportional to X_u^TX_ww_q where w_q is a vector representing the text query facet, X_w is a term-document matrix associating word with documents of the database and X_u is a participant-document matrix associating participants with documents of the database.

16. The retrieval method of claim 10, wherein the performing of query enrichment further includes:

constructing a joint topic-community model defining both the topic model and the community model in terms of hypertopics with both the topics of the topic model and the communities of the community model associated with hypertopics.

17. The retrieval method of claim 16, wherein the database of documents comprises an electronic mail (email) mailbox, the receiving comprises receiving the multi-faceted retrieval query from a computer running an email client, and the retrieval method further includes transmitting an output of the performed ranking to the computer running the email client.

18. A non-transitory computer-readable storage medium storing instructions executable by a computer to generate a set of topics comprising word distributions and a set of communities comprising participant distributions by performing the non-negative matrix factorizations:

19. The non-transitory computer-readable storage medium of claim 18 wherein the non-negative matrix factorizations are performed using the probabilistic formulations

20. The non-transitory computer-readable storage medium of claim 18 further storing instructions executable to perform a retrieval operation comprising:

receiving a retrieval query having at least one of a text query facet comprising one or more keywords and a persons query facet comprising one or more participant names; and

performing query enrichment including at least one of (1) a topic expansion of the text query facet based on the set of topics and (2) a community expansion of the persons query facet based on the set of communities.