Centroid guided contrastive loss improves fraud job posting detection
Beyond Accuracy: Centroid-Guided Contrastive Loss for Structured Fraudulent Job Posting Detection
Artificial Intelligence
Summary
Fraudulent job postings are fake or misleading ads that hurt job seekers and employers. The authors developed a new way to train computers to spot these fake ads more accurately while also understanding subtle differences among them. Their approach, Centroid-Guided Contrastive Loss, helps the computer learn better grouping and separation of genuine and fake posts. Tests on a public job posting dataset showed their method outperformed others.
What this means in practice
- •For online job platform teams: Add improved fake job ad detection to protect users and maintain trustworthy listings on hiring websites.
- •For cybersecurity engineers: Use the centroid-guided contrastive loss method to enhance detection systems for subtle fraudulent or misleading postings.
Authors
Syed Ali Ahmed, Malaika Raza, Muhammad Shoaib Siddiqui, Muhammad Rafi
Abstract
Fraudulent job posting detection aims to identify job advertisements that are corrupted either through fake content, misleading information, or negative intent, disrupting the online eco-system of job-seekers and employers. Existing studies in this domain lack effective methods to simultaneously achieve high accuracy and meaningful structure of latent-space representations that capture subtleties among fake posts. To this end, we propose Centroid-Guided Contrastive Loss (CGCL), a loss function which unifies classification with densely formulated clustering to consistently reshape latent-space through a centroid-driven top-$k$ push-and-pull mechanism. The complementary nature of CGCL enables the model to enforce accurate decision boundaries and maintain high clustering compactness, effectively capturing both class separability and latent structure. Extensive experiments demonstrate the state-of-the-art (SOTA) performance of our method on EMSCAD, a public benchmark dataset. The code associated with this work is available at: https://github.com/ali-ahmed925/CGCL_code