Machine learning to infer title levels across entities
AssigneeMicrosoft Technology Licensing, LLC
InventorsHuichao Xue, Xiaoqing Wang, Chao Wang
Life of this patent
Estimating the term…
Abstract
In an example embodiment, machine learning is used to train a machine-learned model that projects each entity, title pair into a single number, called a seniority score, to represent the career progression needed for that position. For example, company A's “software engineer” and company B's “senior software engineer” can be represented as two separate numbers, one being p (company A, software engineer) and the other being p (company B, senior software engineer) on the same axis. This allows a comparison to be made about the absolute levels of each title despite their potential different meanings at different entities.
Claims20
A system, comprising:
- circuitry;
- and
- a memory to store instructions that, when executed by the circuitry, causes the circuitry to:
- access training data including information about job transitions of users, a job transition including information about an employing entity and a job title for a source job and an employing entity and job title for a target job, and a duration of time between a start of the source job and a start of the target job;
- transform the training data so that vectors in the training data correspond to different job transitions in the training data, wherein a vector comprises a standardized format of a series of triples, each triple comprising an entity identifier, a job title, and a duration of time associated with the job title;
- and
- train an embedding model to calculate a velocity for a vector of a transition in the training data, the velocity comprising a seniority score for an employing entity and a job title for a target job of the transition minus a seniority score for an employing entity and job title for a source job of the given transition, divided by the duration of time for the transition, the trained embedding model to predict a seniority score for a first job having a first employing entity and a first job title.
A method, comprising:
- accessing training data including information about job transitions of users, a job transition including information about an employing entity and a job title for a source job and an employing entity and job title for a target job, and a duration of time between a start of the source job and a start of the target job;
- transforming the training data so that vectors in the training data correspond to different job transitions in the training data, wherein a vector comprises a standardized format of a series of triples, each triple comprising an entity identifier, a job title, and a duration of time associated with the job title;
- and
- training an embedding model to calculate a velocity for a vector of a transition in the training data, the velocity comprising a seniority score for an employing entity and a job title for a target job of the transition minus a seniority score for an employing entity and job title for a source job of the given transition, divided by the duration of time for the transition, the trained embedding model to predict a seniority score for a first job having a first employing entity and a first job title.
A non-transitory machine-readable storage including machine-readable instructions, when executed by circuitry, causes the circuitry to:
- access training data including information about job transitions of users, a job transition including information about an employing entity and a job title for a source job and an employing entity and job title for a target job, and a duration of time between a start of the source job and a start of the target job;
- transform the training data so that vectors in the training data correspond to different job transitions in the training data, wherein a vector comprises a standardized format of a series of triples, each triple comprising an entity identifier, a job title, and a duration of time associated with the job title;
- and
- train an embedding model to calculate a velocity for a vector of a transition in the training data, the velocity comprising a seniority score for an employing entity and a job title for a target job of the transition minus a seniority score for an employing entity and job title for a source job of the given transition, divided by the duration of time for the transition, the trained embedding model to predict a seniority score for a first job having a first employing entity and a first job title.
Dependencies are inferred from the claim text.