Context-based search query formation

Patent metadata

Field	Value
Publication number	US-10984337-B2
Application number	US-201213408853-A
Country	US
Kind code	B2
Filing date	Feb 29, 2012
Priority date	Feb 29, 2012
Publication date	Apr 20, 2021
Grant date	Apr 20, 2021

Abstract

Searching is assisted by recognizing a selection of text from a document as an indication that a user wishes to initiate a search based on the selected text. The user is provided with query suggestions based on the selected text and the query suggestions are ranked based on a context provided by the document. The user may select the text by using a mouse, drawing a circle around the text on a touch screen, or by other input techniques. The query suggestions may be based on query reformulation or query expansion techniques applied to the selected text. Context provided by the document is used by a language model and/or an artificial intelligence system to rank the query suggestions in predicted order of relevance based on the selected text and the context.

First claim

The invention claimed is:

1. An information-processing system comprising:
one or more processing elements;
a search initiation module communicatively coupled to or integrated with the one or more processing elements, the search initiation module configured to:
receive an input indicating text selected in a user interface for presenting a document displayed in a device; and
determine a context of the selected text, the context including a portion of the document that is additional text separate from the selected text;

a candidate query generator coupled to or integrated with the one or more processing elements and configured to identify a plurality of candidate queries based on the selected text, the determined context, and search queries generated by users after viewing the document; and
a query ranking module coupled to or integrated with the one or more processing elements and configured to:
compare each candidate query of the plurality of candidate queries to the selected text and the context;
determine a value associated with each candidate query based on the selected text, and the determined context, the value representing a likelihood that the candidate query corresponds to the selected text and the determined context; and

rank the plurality of candidate queries based at least in part on the value, wherein one or more of the candidate queries are presented on the user interface while the document is being presented based on the determined values.

2. The information-processing system of claim 1 , wherein the candidate query generator is further configured to include, in one or more candidate queries of the plurality of candidate queries, at least one of synonyms of words in the selected text, alternate morphological forms of words in the selected text, correct spellings of misspelled words in the selected text, or alternative spellings of words in the selected text.

3. The information-processing system of claim 1 , wherein the value for each candidate query is determined by an artificial intelligence system, wherein the artificial intelligence system is trained with training data comprising a history of searches originated by users when browsing a corpus of documents, the history of searches being labeled by human labelers indicating a probability that content of a document caused a respective user to submit the corresponding query in the history.

4. A method comprising:
receiving, in a user interface for presenting a document, a selection of text in the document;
determining a context of the selection, the context including additional text from the document that is relative to, and separate from, the selection;
generating a plurality of candidate queries that includes queries generated at least in part by applying one or more query expansion techniques to the text and the context;
comparing each candidate query of the plurality of candidate queries to the text and the context;
determining a value for each candidate query based on the selection of text, and the determined context, the value representing a likelihood that the candidate query corresponds to the selection of text and the determined context;
ranking, by one or more processing elements, the plurality of candidate queries based at least in part on the values;
presenting the plurality of candidate queries in a list ordered at least partly according to the ranking, wherein one or more of the candidate queries are presented on a user interface while the document is being presented based on the determined values;
receiving, in the user interface, a selection of one of the presented candidate queries; and
submitting the selected candidate query to a search engine.

5. The method of claim 4 , wherein the document comprises a mark-up language document.

6. The method of claim 4 , wherein the plurality of candidate queries includes at least one pre-formulated query associated with the document.

7. The method of claim 4 , wherein the one or more query expansion techniques comprise at least one of applying a K-means algorithm to a query log, conducting a random walk on a bipartite query-document graph generated by parsing a query log, running a PageRank algorithm on a query-flow graph generated from a query log, or mining term association patterns from a query log.

8. The method of claim 4 , wherein the additional text comprises at least part of a paragraph of the document, at least part of a column of the document, at least part of a sentence of the document, at least part of a cell of the document, or at least part of a frame of the document.

9. The method of claim 4 , wherein:
ranking the plurality of candidate queries is further based on a language model; and
the language model is based at least in part on a number of words in the candidate query, a number of words in the text, and a number of words in the context.

10. The method of claim 4 , wherein:
ranking the plurality of candidate queries is further based on a language model; and
the language model comprises a bi-gram language model in which a word in the candidate query depends on an immediately preceding word in the candidate query.

11. The method of claim 4 , wherein:
ranking the plurality of candidate queries is further based on an artificial intelligence system; and
the artificial intelligence system learns a function that predicts a level of confidence in one or more candidate queries of the plurality of candidate queries given the candidate query, the selection of text, and the context.

12. The information-processing system of claim 1 , wherein the portion of the document spans at least one sentence, one paragraph, or one column of the document.

13. The information-processing system of claim 1 , wherein the portion of the document includes the selected text.

14. One or more computer storage media, wherein the one or more computer storage media is at least one device, having computer-executable instructions which, when executed by a processor, cause a computing system to:
receive, in a user interface for presenting a document, a selection of text in the document;
determine a context of the selection, the context including additional text from the document that is relative to, and separate from, the selection;
interpret the selection of the text as a command to provide one or more search queries based at least in part on the text;
generate a plurality of candidate queries based at least in part on the text;
compare each candidate query of the plurality of candidate queries to the text and the context;
determine a value for each candidate query based on the selection of text, and the determined context, the value representing a likelihood that the candidate query corresponds to the selection of text and the determined context;
rank the plurality of candidate queries based at least in part on the values to determine a ranking of the plurality of candidate queries;
present, on a user interface while the document is being presented, a subset of the plurality of candidate queries in a list ordered at least partly according to the ranking;
receive, in the user interface, a selection of a candidate query of the subset of the plurality of candidate queries; and
submit the selected candidate query to a search engine.

15. The one or more computer storage media of claim 14 , wherein the computer-executable instructions, when executed by the processor, further cause the computing system to receive the selection of the text based at least in part on a user dragging a pointing implement across the text that is displayed on a touch-screen display.

16.

Assignees

• Bai Peng
• Chen Zheng
• Huang Xuedong David
• Ni Xiaochuan
• Sun Jian Tao
• Zhang Zhimin
• Microsoft Technology Licensing Llc

Inventors

• Bai Peng
• Chen Zheng
• Huang Xuedong David
• Ni Xiaochuan
• Sun Jian-Tao
• Zhang Zhimin

Classifications

• G06F16/3322 Primary

  using system suggestions (G06F16/3325 takes precedence) · CPC title

• G06N20/00 Primary

  Machine learning · CPC title