Contents

Labor space

a high-dimensional representation of the labor market via large language models

KIM, Seongwoon

  • 94 ITEM VIEW
  • 64 DOWNLOAD
Abstract

The labor market is a complex ecosystem comprising multiple economic units such as skills, jobs, industries, and firms. Hence, a true understanding of the labor market requires a holistic perspective that considers the interrelationships between these entities. However, existing studies have often focused on single or bipartite units; therefore, they don’t capture the reciprocal effect of heterogeneous units of the labor market. Here, we introduce Labor Space, a high-dimensional space created by a large language model. Labor Space maps industry, firm, occupation, and skill to a unified embedding space representing the conceptual similarity of the labor market entities. Alignment of conceptual dimensions, such as the production-healthcare axis, reveals our numerical representation portrays the industrial structure of the labor market. Moreover, the calculation of the embedding vector catches the latent relationship of the labor market entities and their interactions with external factors, such as the impact of AI on the labor market. Labor Space offers a comprehensive and innovative approach to understanding the interconnectedness of entities within the labor market, providing a pragmatic tool for researchers, policymakers, and business leaders

Advisors
Park, Jaehyuk
Department
KDI School, Master of Public Policy
Issue Date
2023
Publisher
KDI School
Description
Thesis(Master) -- KDI School: Master of Public Policy, 2023
Keywords
Labor market--Econometric models
Outstanding
Outstanding
Contents
1. Introduction
2. Related Works
3. Data and Methods
4. Landscape of Labor Space
5. Mapping heterogeneous units on conceptual axis
6. Vector calculation for economic analogy
7. Estimating the impact of AI
8. Discussion
9. Appendix
10. Reference
Pages
33 p
URI
https://archives.kdischool.ac.kr/handle/11125/51538
Type
Thesis
Files in This Item:

Click the button and follow the links to connect to the full text. (KDI CL members only)

qrcode

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

상단으로 이동