Patents by Inventor Michael J. Witbrock
Michael J. Witbrock 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).
-
Patent number: 11823013Abstract: Embodiments of the present invention provide a computer-implemented method for performing unsupervised feature representation learning for text data. The method generates reference text data having a set of random text sequences, in which each text sequence of set of random text sequences is of a random length and comprises a number of random words, and in which each random length is sampled from a minimum length to a maximum length. The random words of each text sequence in the set are drawn from a distribution. The method generates a feature matrix for raw text data based at least in part on a set of computed distances between the set of random text sequences and the raw text data. The method provides the feature matrix as an input to one or more machine learning models.Type: GrantFiled: August 29, 2017Date of Patent: November 21, 2023Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Michael J. Witbrock, Lingfei Wu
-
Patent number: 11494677Abstract: An example operation may include one or more of connecting to a blockchain containing chains of reasoning data and related premise data, receiving an inference query, retrieving from the blockchain chain of reasoning data and related premise data corresponding to the inference query, executing inference steps using the retrieved chain of reasoning data and the related premise data to generate conclusion data, tracking the execution of the inference steps, and sending to the blockchain the tracked inference steps and the conclusion data to be entered into the blockchain as transactions.Type: GrantFiled: September 5, 2018Date of Patent: November 8, 2022Assignee: International Business Machines CorporationInventor: Michael J. Witbrock
-
Patent number: 11366990Abstract: Embodiments of the present invention provide a computer-implemented method for performing unsupervised time-series feature learning. The method generates a set of reference time-series of random lengths, in which each length is uniformly sampled from a predetermined minimum length to a predetermined maximum length, and in which values of each reference time-series in the set are drawn from a distribution. The method generates a feature matrix for raw time-series data based on a set of computed distances between the generated set of reference time-series and the raw time-series data. The method provides the feature matrix as an input to one or more machine learning models.Type: GrantFiled: May 15, 2017Date of Patent: June 21, 2022Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Michael J. Witbrock, Lingfei Wu, Cao Xiao, Jinfeng Yi
-
Publication number: 20200074330Abstract: An example operation may include one or more of connecting to a blockchain containing chains of reasoning data and related premise data, receiving an inference query, retrieving from the blockchain chain of reasoning data and related premise data corresponding to the inference query, executing inference steps using the retrieved chain of reasoning data and the related premise data to generate conclusion data, tracking the execution of the inference steps, and sending to the blockchain the tracked inference steps and the conclusion data to be entered into the blockchain as transactions.Type: ApplicationFiled: September 5, 2018Publication date: March 5, 2020Inventor: Michael J. Witbrock
-
Publication number: 20190065986Abstract: Embodiments of the present invention provide a computer-implemented method for performing unsupervised feature representation learning for text data. The method generates reference text data having a set of random text sequences, in which each text sequence of set of random text sequences is of a random length and comprises a number of random words, and in which each random length is sampled from a minimum length to a maximum length. The random words of each text sequence in the set are drawn from a distribution. The method generates a feature matrix for raw text data based at least in part on a set of computed distances between the set of random text sequences and the raw text data. The method provides the feature matrix as an input to one or more machine learning models.Type: ApplicationFiled: August 29, 2017Publication date: February 28, 2019Inventors: Michael J. Witbrock, Lingfei Wu
-
Publication number: 20180330201Abstract: Embodiments of the present invention provide a computer-implemented method for performing unsupervised time-series feature learning. The method generates a set of reference time-series of random lengths, in which each length is uniformly sampled from a predetermined minimum length to a predetermined maximum length, and in which values of each reference time-series in the set are drawn from a distribution. The method generates a feature matrix for raw time-series data based on a set of computed distances between the generated set of reference time-series and the raw time-series data. The method provides the feature matrix as an input to one or more machine learning models.Type: ApplicationFiled: May 15, 2017Publication date: November 15, 2018Inventors: Michael J. Witbrock, Lingfei Wu, Cao Xiao, Jinfeng Yi
-
Patent number: 7228493Abstract: A method includes scanning content of a web page in which a web component is to be inserted, inferring a profile from the scanned content and serving the web component in the web page according to the profile.Type: GrantFiled: March 9, 2001Date of Patent: June 5, 2007Assignee: Lycos, Inc.Inventors: Donald M. Kosak, Michael J. Witbrock
-
Patent number: 6944609Abstract: A method for producing a results list for a search query includes producing a first results list of entries from a search algorithm, the first results list corresponding to a term in the search query, the first results list being ordered according to relevance rankings, modifying a feature of the first results lists by an editor, the modification being made according to an editorial rule, determining a reliability score for the editor, the reliability score being based upon the modification made the editor, combining the modification made by the editor in a combined search index, the combined modification being included in the combined search index according to the determined reliability of the editor, and using the combined search index to produce a second results list which corresponds to the term in the search query, the second results list being ordered according to relevance rankings.Type: GrantFiled: October 18, 2001Date of Patent: September 13, 2005Assignee: Lycos, Inc.Inventor: Michael J. Witbrock
-
Patent number: 6581057Abstract: Disclosed is a computer-assisted method for generating a summary of or a browsing aid for a document. At an index creation time, information that is relevant to at least one dummy query and is necessary to compile at least one temporary summary for the summary or browsing aid is extracted from a document and cached for later use. The information may be compiled into the summary and saved as such. At a search time, the summary or browsing aid is generated using the information that was cached at index creation time. An apparatus for performing this computer-assisted method is also disclosed.Type: GrantFiled: May 9, 2000Date of Patent: June 17, 2003Assignee: Justsystem CorporationInventors: Michael J. Witbrock, Mark Kantrowitz, Vibhu O. Mittal
-
Publication number: 20030078914Abstract: A method for producing a results list for a search query includes producing a first results list of entries from a search algorithm, the first results list corresponding to a term in the search query, the first results list being ordered according to relevance rankings, modifying a feature of the first results lists by an editor, the modification being made according to an editorial rule, determining a reliability score for the editor, the reliability score being based upon the modification made the editor, combining the modification made by the editor in a combined search index, the combined modification being included in the combined search index according to the determined reliability of the editor, and using the combined search index to produce a second results list which corresponds to the term in the search query, the second results list being ordered according to relevance rankings.Type: ApplicationFiled: October 18, 2001Publication date: April 24, 2003Inventor: Michael J. Witbrock
-
Publication number: 20030014501Abstract: A popularity predicting process for determining the popularity of a text-based object includes a query analysis process for analyzing a query to determine a plurality of links to Internet objects relating to the query. A link weighting process determines the individual link strength of each of the plurality of links, thus generating a plurality of link strengths. A link strength summing process determines the sum of the plurality of link strengths, wherein the sum corresponds to the popularity of the text-based object.Type: ApplicationFiled: July 10, 2001Publication date: January 16, 2003Inventors: Andrew R. Golding, Michael J. Witbrock, Alden DoRosario
-
Publication number: 20020129063Abstract: A method includes scanning content of a web page in which a web component is to be inserted, inferring a profile from the scanned content and serving the web component in the web page according to the profile.Type: ApplicationFiled: March 9, 2001Publication date: September 12, 2002Inventors: Donald M. Kosak, Michael J. Witbrock
-
Patent number: 6317708Abstract: A computer method for preparing a summary string from a source document of encoded text. The method comprises comparing a training set of encoded text documents with manually generated summary strings associated therewith to learn probabilities that a given summary word or phrase will appear in summary strings given a source word or phrase appears in encoded text documents and constructing from the source document a summary string containing summary words or phrases having the highest probabilities of appearing in a summary string based on the learned probabilities established in the previous step.Type: GrantFiled: July 12, 1999Date of Patent: November 13, 2001Assignee: JustSystem CorporationInventors: Michael J. Witbrock, Vibhu O. Mittal