System for semantic determination of job titles

Patent metadata

Field	Value
Publication number	US-10860803-B2
Application number	US-201815968751-A
Country	US
Kind code	B2
Filing date	May 2, 2018
Priority date	May 7, 2017
Publication date	Dec 8, 2020
Grant date	Dec 8, 2020

Abstract

A system is described which accepts corporate title and employee data associated with that corporate title data at a first company, putting the corporate title and employee data through a configured network and generating a vector of terms and a set of coefficients associated with that title. Information about an employee is put through a second network using those terms and coefficients to determine if the employee would have the same or similar title at the first company.

First claim

What is claimed is:

1. A system for determining the relationship between two job titles at one or more companies, the system comprising:
a computer,
a title data store, coupled to the computer, configured to accept title, corporate entity, job description and interest data,
a training module, coupled to the computer, configured to describe a neural network, configured to accept input data from the title data store, calculate or generate a set of coefficients and a set of title vectors for the neural network from the input data, and store the coefficients into a database, and
a prediction module, coupled to the computer, configured to accept a first title, a first corporate designation data, a second title and a second corporate designation data, configured to accept the coefficients from the training module, and calculate the similarity between the first title and second title, wherein at least one of the title vectors in the set of title vectors is based on an association of words within the job description in the input data with a weight that is adjusted based on backpropagation as a function of a relationship between said at least one title vector and a known title vector.

2. The system described in claim 1 , further comprising a People Embedding Module, the People Embedding Module coupled to the computer, configured to accept data from the training module and one or more titles and return similar titles.

3. A method for determining the relationship between two job titles at one or more companies, the method using:
a computer,
a title data store, coupled to the computer, configured to accept title, corporate entity, job description and interest data,
a training module, coupled to the computer, configured to describe a neural network, configured to accept input data from the title data store, calculate or generate a set of coefficients and a set of title vectors for the neural network from the input data, and store the coefficients into a database, and
a prediction module, coupled to the computer, configured to accept a first title, a first corporate designation data, a second title and a second corporate designation data, configured to accept the coefficients from the training module, and calculate the similarity between the first title and second title,
the method comprising:
accepting title data from an external data source,
calculating a set of coefficients from said title data,
accepting a title,
processing said title against said set of coefficients, and
returning a set of similar titles,

wherein at least one of the title vectors in the set of title vectors is based on an association of words within the job description in the input data with a weight that is adjusted based on backpropagation as a function of a relationship between said at least one title vector and a known title vector.

4. The system of claim 1 , wherein:
the title data store includes memory circuitry that stores the title, corporate entity, job description and interest data;
the training module includes circuitry and instructions that, when executed by the computer, calculates the set of coefficients for the neural network and stores the coefficients in the database; and
the prediction module includes circuitry and instructions that, when executed by the computer, calculates the similarity between the first title and the second title.

5. The system of claim 1 , wherein the training module is configured to generate the coefficients by creating title vectors based on the input data, each title vector being indicative of a relationship between the title and corporate designation data.

6. The system of claim 5 , wherein the training module is configured to concatenate each title vector with a vector representing the corporate designation data to generate a single concatenated vector represented by the coefficients.

7. The system of claim 5 , wherein the training module is configured to generate the coefficients by, for each title vector:
associating words within the job description in the input data with a weight,
comparing a the title vector to a known title vector, and
adjusting the weight associated with the words via backpropagation, based on the comparing.

8. The system of claim 1 , wherein the prediction module is configured to calculate the similarity between the first title and the second title by determining whether the first title and the second title are semantically similar based on the first title, the second title, and the coefficients.

9. The system of claim 1 , wherein the prediction module is configured and arranged with the training module to calculate the similarity between the first title and the second title by determining a dot product for interests associated with the first title and data associated with the job description and passing the dot product through a sigmoid function to produce a similarity value.

10. The system of claim 1 , wherein the prediction module is configured and arranged with the training module to calculate the similarity between the first title and the second title by:
for each of the first and second titles,
embedding the title using an embedding function and concatenating an output of the embedding function the corporate designation associated with the title, therein producing a single output vector for each title;
predicting a role associated with the title using a classifier; and

comparing the predicted roles for the first and second titles using a similarity function.

11. The system of claim 10 , wherein comparing the predicted roles for the first and second titles using a similarity function includes determining a Euclidian distance between vectors representing the predictions.

12. The system of claim 10 , wherein comparing the predicted roles for the first and second titles using a similarity function includes determining a cosine similarity between the vectors representing the predictions.

13. The system of claim 1 , wherein the prediction module is configured and arranged with the training module to calculate the similarity between the first title and multiple titles, and to return a set of titles from the multiple titles that match the first title.

14. The method of claim 3 , wherein:
the computer includes circuitry;
the title data store includes memory circuitry that stores the title, corporate entity, job description and interest data;
the training module includes circuitry and instructions that, when executed by the computer, calculates the set of coefficients for the neural network and stores the coefficients in the database; and
the prediction module includes circuitry and instructions that, when executed by the computer, calculates the similarity between the first title and the second title.

15. The method of claim 3 , further comprising using the computer, title data store, training module and prediction module to calculate the set of coefficients by:
creating title vectors based on the input data, each title vector being indicative of a relationship between the title and corporate designation data, and
generate a single concatenated vector represented by the coefficients by concatenating each title vector with a vector representing that title vector's corporate designation data.

16. The method of claim 15 , wherein using the computer, title data store, training module and prediction module to calculate the set of coefficients includes, for each title vector:
associating words within the job description in the input data with a weight,
comparing a the title vector to a known title vector, and
adjusting the weight

Assignees

• Fung Solomon
• Mitra Soumyadeb
• Kashyap Abishek
• Samat Arunim
• Nagaswamy Venkat
• Driemeyer Justin
• 8X8 Inc

Inventors

• Fung Solomon
• Mitra Soumyadeb
• Kashyap Abishek
• Samat Arunim
• Nagaswamy Venkat
• Driemeyer Justin

Classifications

• G06F40/30 Primary

  Semantic analysis · CPC title

• G06Q30/0276 Primary

  Advertisement creation · CPC title

• G06F18/21355

  nonlinear criteria, e.g. embedding a manifold in a Euclidean space · CPC title

• G06N3/048

  Activation functions · CPC title

• G06N3/045

  Combinations of networks · CPC title