Should we embed? A study on the online performance of utilizing embeddings for real-time job recommendations
Online job recommendation performance comparison showing Click-Through Rate and Runtime metrics across different approaches, with significance levels indicated by asterisks and relative improvements shown with arrows.
Abstract
In this work, we present the findings of an online study, where we explore the impact of utilizing embeddings to recommend job postings under real-time constraints. On the Austrian job platform Studo Jobs, we evaluate two popular recommendation scenarios - (i) providing similar jobs and, (ii) personalizing the job postings that are shown on the homepage. Our results show that for recommending similar jobs, we achieve the best online performance in terms of Click-Through Rate when we employ embeddings based on the most recent interaction. To personalize the job postings shown on a user’s homepage, however, combining embeddings based on the frequency and recency with which a user interacts with job postings results in the best online performance.
Tomislav Đuričić
Researcher / Machine Learning Engineer / Software Engineer
My research interests include social-based recommender systems, graph neural networks and user modeling.