Le E Zhang

US Patent:

20210019375, Jan 21, 2021

Filed:

Jul 18, 2019

Appl. No.:

16/515318

Inventors:

- Armonk NY, US
Le Zhang - Cary NC, US
Vikrant Verma - Raleigh NC, US
Zhe Zhang - Cary NC, US

International Classification:

G06F 17/28
G06F 17/27
H04L 12/58

Abstract:

A computer-implemented method of executing a virtual agent bot includes receiving, via a computer server, at least one input query from a user, and analyzing the at least one input query to extract at least one input term. The method further comprises determining a semantic skill set of the virtual agent bot among a plurality of different candidate skill sets based at least in part on the at least one input term; and invoking the virtual agent bot to provide a semantic topic response corresponding to the semantic skill set to provide an answer to the at least one input query.

US Patent:

20210012217, Jan 14, 2021

Filed:

Jul 10, 2019

Appl. No.:

16/508000

Inventors:

- Armonk NY, US
Zhe Zhang - Cary NC, US
Le Zhang - Cary NC, US
Vikrant Verma - Raleigh NC, US

International Classification:

G06N 5/02
G06F 16/2455

Abstract:

An approach is provided that receives a question at a question-answering (QA) system. A number of passages are identified that are relevant to the received question. A question knowledge graph is generated that corresponds to the question and a set of passage knowledge graphs are also generated with each passage knowledge graph corresponding to one of the identified passages. Each of the passage knowledge graphs are compared to the question knowledge graph with the comparison resulting in a set of knowledge graph candidate answers (kgCAs). A set of candidate answers (CAs) is computed by the QA with at least one of the CAs being based on one of the kgCAs.

US Patent:

20210012218, Jan 14, 2021

Filed:

Jul 10, 2019

Appl. No.:

16/508038

Inventors:

- Armonk NY, US
Zhe Zhang - Cary NC, US
Le Zhang - Cary NC, US
Vikrant Verma - Raleigh NC, US

International Classification:

G06N 5/02
G06F 17/27
G06F 16/30

Abstract:

An approach is provided that selects an original entity from an original knowledge graph. The approach then accesses a data source that is external to the original knowledge graph, such as an online encyclopedia. An entity in the data source is identified based on the entity matching the original entity. A new relation is then identified in the data source between the identified entity and a new entity with the new entity being absent from the original knowledge graph. An expanded knowledge graph is then generated with the expanded knowledge graph formed by adding the new entity to the original knowledge graph.

US Patent:

20200110834, Apr 9, 2020

Filed:

Oct 8, 2018

Appl. No.:

16/154057

Inventors:

- Armonk NY, US
Vikrant Verma - Raleigh NC, US
Zhe Zhang - Cary NC, US
Le Zhang - Cary NC, US

Assignee:

International Business Machines Corporation - Armonk NY

International Classification:

G06F 17/30
G06N 99/00

Abstract:

Embodiments are directed to a system, computer program product, and method for identification of linguistically related elements, and more specifically to prediction of a linguistically related element. A linguistic algorithm forms a cluster representation of corpus entries. A linguistic term is identified and application to the cluster representation to identified proximally related linguistic terms. Associative relationships between the proximally related terms and category metadata are iteratively investigated. One or more linguistic terms related across the two more metadata categories is identified and designated at the linguistically related element.

US Patent:

20180232373, Aug 16, 2018

Filed:

Feb 13, 2017

Appl. No.:

15/430597

Inventors:

- Armonk NY, US
Vikrant Verma - Cary NC, US
Le Zhang - Durham NC, US
Zhe Zhang - Cary NC, US

International Classification:

G06F 17/30

Abstract:

An approach is provided that receives a question at a question answering (QA) system. The question includes a number of words. The approach operates by calculating weights that correspond to search terms included in the plurality of words. The search terms include the plurality of words and may include terms that are one or more sequences of adjacent words included in the question. Based on the calculated weights and the words in the question, the approach generates a query that is used to search a corpus that is managed by the QA system with the search resulting in one or more search results.

US Patent:

20180232374, Aug 16, 2018

Filed:

Jun 12, 2017

Appl. No.:

15/619689

Inventors:

- Armonk NY, US
Vikrant Verma - Cary NC, US
Le Zhang - Durham NC, US
Zhe Zhang - Cary NC, US

International Classification:

G06F 17/30

Abstract:

An approach is provided that receives a question at a question answering (QA) system. The question includes a number of words. The approach operates by calculating weights that correspond to search terms included in the plurality of words. The search terms include the plurality of words and may include terms that are one or more sequences of adjacent words included in the question. Based on the calculated weights and the words in the question, the approach generates a query that is used to search a corpus that is managed by the QA system with the search resulting in one or more search results.

US Patent:

20180089305, Mar 29, 2018

Filed:

Sep 28, 2016

Appl. No.:

15/278151

Inventors:

- Armonk NY, US
VIKRANT VERMA - CARY NC, US
LE ZHANG - DURHAM NC, US
ZHE ZHANG - CARY NC, US

International Classification:

G06F 17/30
G06F 17/27

Abstract:

A computer-implemented method generates a candidate answer triple for use in retrieving information used to answer a question. One or more processors parse a question to identify a lexical answer type for the question, a question action for the question, and a question timestamp for the question to make up a question triple. One or more processors retrieve multiple candidate passages for answering the question, and parse each of the multiple candidate passages to identify a candidate entity, a candidate action, and a candidate timestamp from each of the multiple candidate passages to generate a candidate answer triple. One or more processors compare the question triple to the candidate answer triple and establish a match score for each candidate answer triple, which is used in retrieving information used to answer the question.

US Patent:

20180089569, Mar 29, 2018

Filed:

Sep 28, 2016

Appl. No.:

15/278172

Inventors:

- Armonk NY, US
VIKRANT VERMA - CARY NC, US
LE ZHANG - DURHAM NC, US
ZHE ZHANG - CARY NC, US

International Classification:

G06N 5/02
G06F 17/30

Abstract:

A computer-implemented method generates a temporal candidate answer to a question. One or more processors tag terms found in multiple text passages with a part of speech (POS) tag, and then parse POS tagged terms found in the multiple text passages to generate passage triples for multiple parts of each of the multiple text passages. The processor(s) filter each sub-passage from the multiple text passages, which removes passages from each of the multiple text passages that are not associated with said each sub-passage. The processor(s) temporally align each sub-passage triple with a particular time range, and then extract text within each sub-passage from the multiple text passages to create a temporal candidate answer identified by the associated sub-passage triple. Upon receipt of a question related to a sub-passage that is related to the particular time range, the processor(s) return the temporal candidate answer that is temporally aligned with the question.