OKRA: an Explainable, Heterogeneous, Multi-Stakeholder Job Recommender System
Roan Schellingerhout, Francesco Barile, Nava Tintarev
TL;DR
OKRA addresses the high-stakes, multi-stakeholder nature of job recommendations by integrating structured and unstructured data into a heterogeneous knowledge graph and learning stakeholder-specific embeddings with attention-based graph networks. The method computes candidate- and company-side scores and combines them via a harmonic mean to produce explainable, ground-grounded rankings, achieving state-of-the-art $nDCG$ on two datasets. However, the study also reveals urban bias in both user and provider exposures, indicating that higher accuracy may come with fairness trade-offs. The work highlights the practicality of explainability and multi-stakeholder grounding in recruitment, while outlining ethical considerations and avenues for fairness-aware improvements and stakeholder-focused evaluation.
Abstract
The use of recommender systems in the recruitment domain has been labeled as 'high-risk' in recent legislation. As a result, strict requirements regarding explainability and fairness have been put in place to ensure proper treatment of all involved stakeholders. To allow for stakeholder-specific explainability, while also handling highly heterogeneous recruitment data, we propose a novel explainable multi-stakeholder job recommender system using graph neural networks: the Occupational Knowledge-based Recommender using Attention (OKRA). The proposed method is capable of providing both candidate- and company-side recommendations and explanations. We find that OKRA performs substantially better than six baselines in terms of nDCG for two datasets. Furthermore, we find that the tested models show a bias toward candidates and vacancies located in urban areas. Overall, our findings suggest that OKRA provides a balance between accuracy, explainability, and fairness.