Patents by Inventor Kin Fai Kan
Kin Fai Kan has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Publication number: 20230259824Abstract: A set of data for training a machine learning system can be modified to improve its performance. An item of information can be transmitted. A message can be transmitted that includes an explanation of a determination, by the machine learning system, to transmit the item of information from among a plurality of items of information. A first set of data can have been used to train the machine learning system. A signal can be received that includes an indication of a usefulness of the message, to a user of a user device, in making a decision to perform an action based on a knowledge associated with the item of information. The first set of data can be modified, in response to a receipt of the signal, to produce a second set of data. The machine learning system can be caused to be trained using the second set of data.Type: ApplicationFiled: April 27, 2023Publication date: August 17, 2023Inventors: Mayukh Bhaowal, Leah McGuire, Kin Fai Kan, Christopher Rupley, Xiaodan Sun, Michael Weil, Subha Nabar
-
Patent number: 11720825Abstract: The system and methods of the disclosed subject matter provide an experimentation framework to allow a user to perform machine learning experiments on tenant data within a multi-tenant database system. The system may provide an experimental interface to allow modification of machine learning algorithms, machine learning parameters, and tenant data fields. The user may be prohibited from viewing any of the tenant data or may be permitted to view only a portion of the tenant data. Upon generating an experimental model using the experimental interface, the user may view results comparing the performance of the experimental model with a current production model.Type: GrantFiled: January 31, 2019Date of Patent: August 8, 2023Assignee: Salesforce, Inc.Inventors: Sarah Aerni, Luke Sedney, Kin Fai Kan, Till Christian Bergmann
-
Patent number: 11669767Abstract: A set of data for training a machine learning system can be modified to improve its performance. An item of information can be transmitted. A message can be transmitted that includes an explanation of a determination, by the machine learning system, to transmit the item of information from among a plurality of items of information. A first set of data can have been used to train the machine learning system. A signal can be received that includes an indication of a usefulness of the message, to a user of a user device, in making a decision to perform an action based on a knowledge associated with the item of information. The first set of data can be modified, in response to a receipt of the signal, to produce a second set of data. The machine learning system can be caused to be trained using the second set of data.Type: GrantFiled: August 15, 2019Date of Patent: June 6, 2023Assignee: Salesforce, Inc.Inventors: Mayukh Bhaowal, Leah McGuire, Kin Fai Kan, Christopher Rupley, Xiaodan Sun, Michael Weil, Subha Nabar
-
Patent number: 11663517Abstract: A system may automatically generate a predictive machine learning model by automatically performing various processes based on an analysis of the data as well as metadata associated with the data. The system may accept a selection of data and a prediction field from the data. The system may automatically generate a set of features based on the data and may automatically remove certain features that cause inaccuracies in the model. The system may balance the data based on a representation rate of certain outcomes. The system may train and select a model based on several candidate models. The system may then perform the predictions based on the selected model and send an indication of the predictions to a user.Type: GrantFiled: January 31, 2018Date of Patent: May 30, 2023Assignee: Salesforce, Inc.Inventors: Sara Beth Asher, John Emery Ball, Vitaly Gordon, Till Christian Bergmann, Kin Fai Kan, Chalenge Masekera, Shubha Nabar, Nihar Dandekar, James Reber Lewis
-
Publication number: 20230110057Abstract: A method for generating a model for recommendations from an item data set for a target data set includes embedding a set of targets from the target data set in a shared coordinate space using a first embedding function, embedding a first set of items from the item data set in the shared coordinate space using a second embedding function, selecting at least one target from the set of targets, and identifying a second set of items from the first set of items that are proximate to the at least one target as candidates from the recommendations.Type: ApplicationFiled: October 7, 2021Publication date: April 13, 2023Applicant: salesforce.com, inc.Inventors: Kin Fai Kan, Chaney Lin, Mayukh Bhaowal, Shubha Nabar, Seiji J. Yamamoto
-
Publication number: 20210049419Abstract: A set of data for training a machine learning system can be modified to improve its performance. An item of information can be transmitted. A message can be transmitted that includes an explanation of a determination, by the machine learning system, to transmit the item of information from among a plurality of items of information. A first set of data can have been used to train the machine learning system. A signal can be received that includes an indication of a usefulness of the message, to a user of a user device, in making a decision to perform an action based on a knowledge associated with the item of information. The first set of data can be modified, in response to a receipt of the signal, to produce a second set of data. The machine learning system can be caused to be trained using the second set of data.Type: ApplicationFiled: August 15, 2019Publication date: February 18, 2021Inventors: Mayukh Bhaowal, Leah McGuire, Kin Fai Kan, Christopher Rupley, Xiaodan Sun, Michael Weil, Subha Nabar
-
Publication number: 20200250587Abstract: The system and methods of the disclosed subject matter provide an experimentation framework to allow a user to perform machine learning experiments on tenant data within a multi-tenant database system. The system may provide an experimental interface to allow modification of machine learning algorithms, machine learning parameters, and tenant data fields. The user may be prohibited from viewing any of the tenant data or may be permitted to view only a portion of the tenant data. Upon generating an experimental model using the experimental interface, the user may view results comparing the performance of the experimental model with a current production model.Type: ApplicationFiled: January 31, 2019Publication date: August 6, 2020Inventors: Sarah Aerni, Luke Sedney, Kin Fai Kan, Till Christian Bergmann
-
Patent number: 10552753Abstract: Techniques for inferring the identity (e.g., member profile attributes) of members of an online social network service are described. According to various embodiments, a member profile attribute missing from a member profile page associated with a particular member of an online social network service is identified. Member profile data and behavioral log data associated with a plurality of members of the online social network service is then accessed. Thereafter, a prediction modeling process is performed, based on a prediction model and feature data including the member profile data and the behavioral log data, to generate a confidence score associated with the particular member and the missing member profile attribute, the confidence score indicating a likelihood that the missing member profile attribute corresponds to a candidate value.Type: GrantFiled: December 10, 2015Date of Patent: February 4, 2020Assignee: Microsoft Technology Licensing, LLCInventors: Zhigang Hua, Kin Fai Kan, Peter N. Skomoroch, Gloria Lau, Saveliy Uryasev
-
Publication number: 20190138946Abstract: A system may automatically generate a predictive machine learning model by automatically performing various processes based on an analysis of the data as well as metadata associated with the data. The system may accept a selection of data and a prediction field from the data. The system may automatically generate a set of features based on the data and may automatically remove certain features that cause inaccuracies in the model. The system may balance the data based on a representation rate of certain outcomes. The system may train and select a model based on several candidate models. The system may then perform the predictions based on the selected model and send an indication of the predictions to a user.Type: ApplicationFiled: January 31, 2018Publication date: May 9, 2019Inventors: Sara Beth Asher, John Emery Ball, Vitaly Gordon, Till Christian Bergmann, Kin Fai Kan, Chalenge Masekera, Shubha Nabar, Nihar Dandekar, James Reber Lewis
-
Patent number: 9886498Abstract: A title standardization system is may be configured to detect an edit operation associated with the job title field of a member profile stored by an on-line social network system and, in response, perform operations to derive a canonical title that represents a raw title string found in the job title field. The derived canonical title may be then associated with the member profile, in which the originally-obtained subject title string was found. This association may be stored in a database for future use, e.g., for targeting job recommendations, recruiting, making professional contacts, as well as for other purposes.Type: GrantFiled: October 24, 2014Date of Patent: February 6, 2018Assignee: Microsoft Technology Licensing, LLCInventors: Arpit Amar Goel, Uri Merhav, Vitaly Gordon, Kin Fai Kan, Craig Martell
-
Publication number: 20160196266Abstract: In order to determine seniority associated with a title string associated with a member profile in an on-line social network system, a standardization system may be configured to operate as follows. A standardization system may determine a canonical title that corresponds to the title string, determine any seniority modifiers that may be present in the title string, and calculate a seniority value for the title sting as the sum of the seniority value assigned to the determined canonical title and the respective seniority values of the determined seniority modifiers. A seniority modifier is a phrase comprising one or more words that have been identified as being indicative of seniority if included in a title string.Type: ApplicationFiled: January 2, 2015Publication date: July 7, 2016Inventors: Uri Merhav, Vitaly Gordon, Kin Fai Kan
-
Publication number: 20160196272Abstract: A title standardization system may be configured to automatically identify modifier terms in title strings and store these terms in a dictionary for future use. Modifier terms are those phrases in a title string that have been identified as indicative of a certain aspect related to the job of the associated member. In order to identify modifier terms, a title standardization system examines transitions between jobs that the members of the on-line social network system have reported via their respective profiles. If a transition pair comprising a first title string and a second title string was determined to be conforming to a stable pattern, a phrase that is included in the first title string and is lacking from the second title string is identified as a modifier phrase and stored in a dictionary for future use.Type: ApplicationFiled: January 2, 2015Publication date: July 7, 2016Inventors: Uri Merhav, Vitaly Gordon, Kin Fai Kan
-
Publication number: 20160196619Abstract: A seniority standardization system may be configured to derive seniority values in the context of an on-line social network system. In order to determine a seniority rank of a given professional title, a seniority standardization system may leverage transition data, which is information that may be gleaned from a member profile with respect to the member's transition from one professional position to another. A seniority standardization system may also use time-based seniority signal. A time-based seniority value, which may be assigned to a particular professional title, is the amount of time that it typically takes to achieve a professional position represented by that particular professional title.Type: ApplicationFiled: January 2, 2015Publication date: July 7, 2016Inventors: Uri Merhav, Vitaly Gordon, Kin Fai Kan
-
Publication number: 20160117385Abstract: A title standardization system is may be configured to detect an edit operation associated with the job title field of a member profile stored by an on-line social network system and, in response, perform operations to derive a canonical title that represents a raw title string found in the job title field. The derived canonical title may be then associated with the member profile, in which the originally-obtained subject title string was found. This association may be stored in a database for future use, e.g., for targeting job recommendations, recruiting, making professional contacts, as well as for other purposes.Type: ApplicationFiled: October 24, 2014Publication date: April 28, 2016Inventors: Arpit Amar Goel, Uri Merhav, Vitaly Gordon, Kin Fai Kan, Craig Martell
-
Publication number: 20160098644Abstract: Techniques for inferring the identity (e.g., member profile attributes) of members of an online social network service are described. According to various embodiments, a member profile attribute missing from a member profile page associated with a particular member of an online social network service is identified. Member profile data and behavioral log data associated with a plurality of members of the online social network service is then accessed. Thereafter, a prediction modeling process is performed, based on a prediction model and feature data including the member profile data and the behavioral log data, to generate a confidence score associated with the particular member and the missing member profile attribute, the confidence score indicating a likelihood that the missing member profile attribute corresponds to a candidate value.Type: ApplicationFiled: December 10, 2015Publication date: April 7, 2016Inventors: Zhigang Hua, Kin Fai Kan, Peter N. Skomoroch, Gloria Lau, Saveliy Uryasev
-
Publication number: 20160026632Abstract: Method and system to infer seniority level of a social network member is provided. The system includes a transition data extractor, a token extractor, a transition data analyzer, and a storing module. The transition data extractor extracts transition data from member profiles maintained in an online social network system. The token extractor extracts a plurality of tokens from title strings in the transition data. The transition data analyzer analyzes the transition data to generate a weight for each token in the plurality of tokens. A weight for a token in the plurality of tokens indicates a contribution of the token to a seniority rank of a title string that includes the token. The storing module stores the plurality of tokens and their associated weights in a database.Type: ApplicationFiled: July 23, 2014Publication date: January 28, 2016Inventors: Gloria Lau, Vitaly Gordon, Kin Fai Kan
-
Publication number: 20150379112Abstract: Method and system to create a job function ontology may be utilized to derive, from member profiles maintained in an on-line social networking system, job function entities associated with respective sets of professional attributes. An entry in the job function ontology—a job function entity—may include identification of the associated job function, as well as a set of professional attributes that characterize professional skills of a member of the on-line social network system. A label assigned to a job function entity may be viewed as a standardized job title.Type: ApplicationFiled: June 27, 2014Publication date: December 31, 2015Inventors: Gloria Lau, Vitaly Gordon, Kin Fai Kan
-
Patent number: 8732014Abstract: A system and method for automatically classifying ads into a taxonomy of categories, the method including: extracting text features from ad images using OCR (optical character recognition) techniques; identifying objects of interest from ad images using object detection and recognition techniques in computer vision; extracting text features from the web-page of the advertiser to which the user is re-directed when clicking the ad; training statistical models using the extracted features mentioned above as well as advertiser attributes from a historical dataset of ads labeled by human editors; and determining the relevant categories of unlabeled ads using the trained models.Type: GrantFiled: December 20, 2010Date of Patent: May 20, 2014Assignee: Yahoo! Inc.Inventors: Andrew Kae, Kin Fai Kan, Vijay K. Narayanan
-
Publication number: 20120158525Abstract: A system and method for automatically classifying ads into a taxonomy of categories, the method including: extracting text features from ad images using OCR (optical character recognition) techniques; identifying objects of interest from ad images using object detection and recognition techniques in computer vision; extracting text features from the web-page of the advertiser to which the user is re-directed when clicking the ad; training statistical models using the extracted features mentioned above as well as advertiser attributes from a historical dataset of ads labeled by human editors; and determining the relevant categories of unlabeled ads using the trained models.Type: ApplicationFiled: December 20, 2010Publication date: June 21, 2012Applicant: Yahoo! Inc.Inventors: Andrew Kae, Kin Fai Kan, Vijay K. Narayanan