Guangsheng Zhang

US Patent:

8370129, Feb 5, 2013

Filed:

Oct 4, 2009

Appl. No.:

12/573134

Inventors:

Guangsheng Zhang - Palo Alto CA, US

International Classification:

G06F 17/27

US Classification:

704 9

Abstract:

A method is disclosed for quantitatively assessing information in natural language contents related to an object name. The method includes identifying a sentence in a document, determining a subject and a predicate in the sentence, and retrieving an object-specific data set related to the object name. The object-specific data set includes property names and association-strength values. Each property name is associated with an association-strength value. The method also includes identifying a first property name in the property names that matches the subject, assigning a first association-strength value associated with the first property name to the subject, identifying a second property name in the property names that matches the predicate, assigning a second association-strength value associated with the second property name to the predicate, and multiplying the first association-strength value and the second association-strength value to produce a sentence information index.

US Patent:

8370347, Feb 5, 2013

Filed:

Feb 17, 2012

Appl. No.:

13/399050

Inventors:

Guangsheng Zhang - Palo Alto CA, US

International Classification:

G06F 7/08
G06F 17/30

US Classification:

707730, 707728, 707750

Abstract:

A system is described for assessing information in natural language contents. A user interface receives an object name as a query term and a value for a customized ranking parameter from a user. A computer storage device stores an object-specific data set related to the object name, wherein the object-specific data set includes a plurality of property names and association-strength values. A computer processing system can count a first frequency of a first property name and count a second frequency of a second property name in a document containing text in a natural language, calculate a relevance score as a function of the first frequency and the second frequency, and rank the plurality of documents using their respective relevance scores, and return one or more documents to the user based on the ranking of the plurality of documents. The function is in part defined by the customized ranking parameter.

US Patent:

8380489, Feb 19, 2013

Filed:

Dec 5, 2009

Appl. No.:

12/631829

Inventors:

Guangsheng Zhang - Palo Alto CA, US

International Classification:

G06F 17/27

US Classification:

704 4

Abstract:

A method for quantitatively assessing associations between different terms in a natural language includes obtaining a first group of sentences, at least some of which comprise an object name, parsing each of the first group of sentences to identify a subject and a predicate, assigning a sentence type to each of the first group of sentences according to the location of the object name in the sentence, tokenizing each of the first group of sentences to produce a plurality of tokens that includes a jth token, and adding a weighting coefficient to a parameter token_j_count for each sentence that includes the jth token, wherein the weighting coefficient is dependent on the sentence type. A cumulative value of the parameter token_j_count obtained from the first group of sentences is divided by the total number of sentences in the first group to produce an internal association strength for the jth token.

US Patent:

8407217, Mar 26, 2013

Filed:

Dec 18, 2010

Appl. No.:

12/972462

Inventors:

Guangsheng Zhang - Palo Alto CA, US

International Classification:

G06F 17/30

US Classification:

707731, 707708, 707723, 707748, 707771

Abstract:

A computer-assisted method for discovering topics in a document includes the steps of calculating a term prominence value for a token, selecting tokens based on the term prominence values of the respective tokens, and outputting topic terms comprising the selected tokens to represent the topics in the document.

US Patent:

8498983, Jul 30, 2013

Filed:

Sep 23, 2011

Appl. No.:

13/241534

Inventors:

Guangsheng Zhang - Palo Alto CA, US

International Classification:

G06F 17/30

US Classification:

707731, 707706, 707722, 707736, 706 12, 706 14

Abstract:

A computer-assisted method for assisting a user to search for documents or other file objects includes receiving a query comprising a queried term from a user, wherein the queried term comprises a sequence of characters entered by the user, obtaining a first context term list comprising a first context term semantically related to the queried term, obtaining and displaying a first search result, dynamically displaying the first context term list in response to each character being entered in the query, allowing the user to select the first context term in the first context term list, obtaining a second context term list comprising a second context term semantically related to the queried term and the first context term, and displaying a second search result and the second context term list in a user interface in response to the selection of the first context term in the first context term list.

US Patent:

20140046976, Feb 13, 2014

Filed:

Mar 15, 2013

Appl. No.:

13/844143

Inventors:

Guangsheng Zhang - Palo Alto CA, US
Chizhong Zhang - Palo Alto CA, US

International Classification:

G06F 17/30

US Classification:

707772

Abstract:

Systems and methods are disclosed for presenting information related to an object including a word or phrase in a text content that is considered as unstructured data. Methods include identifying a term in the text content as a candidate for presenting such information. Once the term is identified, various data sources including document or email search indexes and relational or non-relational databases are searched, and data units related to the selected terms are retrieved and displayed in a user interface with a concurrent view of both the text content containing the term and the related information. Other methods for utilizing various visual effects to concurrently display relevant information associated with various user interface objects are also disclosed. Further methods disclosed include effective ways for handling emails, enhancing social network information presentation and user experiences, as well as enterprise internal message delivery for enterprise healthcare cost-reduction and employee well-being.

US Patent:

8140526, Mar 20, 2012

Filed:

Feb 3, 2010

Appl. No.:

12/699193

Inventors:

Guangsheng Zhang - Palo Alto CA, US

International Classification:

G06F 7/04
G06F 17/30

US Classification:

707730, 707750, 707749, 707752

Abstract:

A system is described for assessing information in natural language contents. A user interface receives an object name as a query term and a value for a customized ranking parameter from a user. A computer storage stores an object-specific data set related to the object name, wherein the object-specific data set includes a plurality of property names and association-strength values. A computer processing system can count a first frequency of a first property name and count a second frequency of a second property name in a document containing text in a natural language, calculate a relevance score as a function of the first frequency and the second frequency, and rank the plurality of documents using their respective relevance scores, and return one or more documents to the user based on the ranking of the plurality of documents. The function is in part defined by the customized ranking parameter.

US Patent:

20140278365, Sep 18, 2014

Filed:

Mar 17, 2013

Appl. No.:

13/844980

Inventors:

Guangsheng Zhang - Palo Alto CA, US
Chizhong Zhang - Palo Alto CA, US

International Classification:

G06F 17/27

US Classification:

704 9

Abstract:

System and methods are disclosed for determining the connotation or sentiment type of a text unit comprising multiple terms and with a grammatical structure, such as subject+verb, verb+object, adjective+noun, noun+noun, noun+preposition+noun. The connotation or sentiment type of the text unit is determined by applying context rules where the context of the grammatical structure may change the inherent or default connotations of individual terms in the text unit. The methods provide a solution to the challenge of correctly or accurately determining the sentiment type of various linguistic structures under different context, and to the simplistic approach of using the inherent or default connotation of individual terms for the linguistic structure containing such terms.