SYSTEMS AND METHODS FOR QUERY REWRITING

• US 20160125028A1
• Filed: 11/05/2014
• Published: 05/05/2016
• Est. Priority Date: 11/05/2014
• Status: Abandoned Application

• Current Assignee: Excalibur IP, LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Silvestri, Fabrizio, Grbovic, Mihajlo, Radosavljevic, Vladan, Djuric, Nemanja, Bhamidipati, Narayan
• Application Number:
  US14/533,405
• Publication Number:
  US 20160125028A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3338 Query expansion

Abstract

Systems and methods for rewriting query terms are disclosed. The system collects queries and query session data and separates the queries into sequences of queries having common sessions. The sequences of queries are then input into a deep learning network to build a multidimensional word vector in which related terms are nearer one another than unrelated terms. An input query is then received and the system matches the input query in the multidimensional word vector and rewrites the query using the nearest neighbors to the term of the input query.

Claims

1. A computing system for rewriting queries, comprising:

an input module configured to receive a plurality of queries and session information for each of the queries;

a learning module configured to embed terms contained in the plurality of queries in a multidimensional word vector, wherein terms having a similar context in a session are near each other in the multidimensional word space; and

a query rewrite module configured to receive a query, find the nearest neighbors of terms within the query in the multidimensional word vector, and rewrite the query with the nearest neighbors of the term.

2. The computing system of claim 1, wherein the plurality of queries and session information comprise at least one document consisting of a string of uninterrupted queries ordered temporally by a user.

3. The computing system of claim 1, wherein the nearest neighbor is found using a cosine distance metric.

4. The computing system of claim 2, wherein a string of uninterrupted queries is defined as an uninterrupted sequence of web search activity by a user that ends when the user is inactive for more than 30 minutes.

5. The computing system of claim 1, wherein input module is further configured to group queries among the plurality of queries into documents comprising a string of uninterrupted queries ordered temporally by a user.

6. The system of claim 1, wherein the learning module operates on the plurality of word sequences in a sliding window fashion.

7. The system of claim 1, wherein each sequence of words is a context.

8. A method for rewriting queries, comprising:

accessing a history of search query activity to obtain a plurality of queries and session data;

grouping queries from among the plurality of queries into documents, with all queries in a document having a common session;

inputting the documents into a deep learning network to embed terms from among the queries in a multidimensional word vector in which related terms are found close to one another;

receiving an input query;

locating terms in the input query within the multidimensional word vector;

finding a plurality of nearest neighbor terms to the input terms in the multidimensional word vector; and

rewriting the input query into a modified query containing the plurality of nearest neighbor terms.

9. The method of claim 8, wherein finding a plurality of nearest neighbor terms comprises determining nearest neighbors through a cosine distance metric.

10. The method of claim 8, wherein the multidimensional word vector has greater than 200 dimensions.

11. A computer program product for rewriting queries, the computer program product comprising non-transient computer readable storage media have instructions stored thereon that cause a computing device to perform a method comprising:

receive a query comprising a query term;

access a multidimensional word vector of interconnected query words to find a plurality of related query words spatially near the query in the multidimensional word vector;

rewrite the query with the plurality of related words.

12. The computer program product of claim 11, wherein the multidimensional word vector comprises an output of a deep learning network trained with a plurality of word sequences with each word sequence comprising query terms from a continuous query session.

13. The computer program product of claim 12, wherein the instructions further cause the computing device to build the multidimensional word vector.

14. The computer program product of claim 13, wherein building the multidimensional word vector comprises:

collecting a plurality of query terms having associated session data;

grouping query terms from among the plurality of query terms according to session data to form term sequences;

inputting the term sequences into a deep learning network to embed each term in a multidimensional word vector in which related terms are found close to one another.

15. The computer program product of claim 11, wherein the query comprises a multi-word phrase.