Use cases

Public Employment

emp

Use Case Description

The purpose of this Use Case is to develop innovative AI approaches for enhancing the efficiency of the job market. This means, methods that ensure that a suitable job is found for each worker, and a suitable worker is found for each job. The fast changes in the job market and the two-sided nature of this problem make it particularly challenging.

Two broad approaches will be developed:

  • In the first approach, job seekers and jobs are assumed fixed, and methods are developed for ensuring an optimal match between them, free from undesirable biases.

  • In the second approach, advice may be given both to job seekers and to employers in how they can reposition themselves to better align with the realities of the current job market. This may be done by re-schooling or upskilling (job seekers) or by adjusting business models or providing in-house training opportunities (employers).

Throughout all this, fairness and transparency will be overarching constraints.

Results, Highlights & Research Outputs

New types of fairness in e-recruitment recommender systems

Tangible results

Several novel formalizations of fairness and related notions in e-recruitment recommender systems, as well as algorithmic approaches for improving those.

This line of work led to a new collaboration with a large Belgian outplacement company.

Publications

  • Mashayekhi, Yoosof, Bo Kang, Jefrey Lijffijt, and Tijl De Bie. "ReCon: Reducing congestion in job recommendation using optimal transport." In Proceedings of the 17th ACM Conference on Recommender Systems, pp. 696-701. 2023.

  • Mashayekhi, Yoosof, Bo Kang, Jefrey Lijffijt, and Tijl De Bie. "GREASE: Graph Imbalance Reduction by Adding Sets of Edges." IEEE Transactions on Knowledge and Data Engineering 36, no. 4 (2023): 1611-1623.

  • Mashayekhi, Yoosof, Bo Kang, Jefrey Lijffijt, and Tijl De Bie. "Quantifying and reducing imbalance in networks." In RECSYS IN HR 2021, vol. 2967. CEUR, 2021.

  • Li, Nan, Bo Kang, Jefrey Lijffijt, and Tijl De Bie. "FEIR: Quantifying and reducing envy and inferiority for fair recommendation of limited resources." ACM Transactions on Intelligent Systems and Technology 15, no. 4 (2024): 1-24.

  • S. Nathan, G. Bied, E. Perennes, P. Caillou, B. Crepon, C. Gaillac, M. Sebag, JoLA: Job Landscape Aware Job Recommendation, forthcoming in the Proceedings of the 5th Workshop on Recommender Systems for Human Resources (RecSys-in-HR 2025), co-located with the 19th ACM Conference on Recommender Systems (RecSys 2025), Prague, Czech Republic, 22nd-26th September 2025

Software / Tool

No toolboxes released, although the code of most of these projects is publicly available.

The public JobHop dataset of career trajectories

Tangible results

A large-scale publicly available dataset of career trajectories, extracted in a privacy-preserving manner from the VDAB database of CVs. An analysis is included as well in the publication.

Publication

  • S. Nathan, G. Bied, E. Perennes, P. Caillou, B. Crepon, C. Gaillac, M. Sebag, JoLA: Job Landscape Aware Job Recommendation, forthcoming in the Proceedings of the 5th Workshop on Recommender Systems for Human Resources (RecSys-in-HR 2025), co-located with the 19th ACM Conference on Recommender Systems (RecSys 2025), Prague, Czech Republic, 22nd-26th September 2025

Software / Tool

  • https://huggingface.co/datasets/aida-ugent/JobHop

Challenge-based survey of e-recruitment recommendation systems

Tangible results

An extensive survey from an innovative perspective of recommendation systems in the job market (e.g. recommending jobs to job seekers, or vice versa).

Publication

  • Mashayekhi, Yoosof, Nan Li, Bo Kang, Jefrey Lijffijt, and Tijl De Bie. "A challenge-based survey of e-recruitment recommendation systems." ACM Computing Surveys 56, no. 10 (2024): 1-33.

Software / Tool

  • Supporting website: https://aida-ugent.github.io/e-recruitment-recsys-challenges/

Inferring missing skills in CVs using PU (positive and unlabeled) learning

Tangible results

A novel method that can be used to infer missing skills in CVs and job seeker profiles based on positive and unlabeled data only, making use of variational inference.

Publication

  • Victor Verreet, Lennert De Smet, Robin Manhaeve, Pieter Delobelle, and Jessa Bekker. "Inferring Missing CV Skills using PU Learning and Variational Inference." In International workshop on AI for Human Resources and Public Employment Services at ECML-PKDD 2023, Date: 2023/09/18-2023/09/22, Location: Turin, Italy. 2023.

Outputs & Resources

Publications

  • Mashayekhi, Yoosof, Bo Kang, Jefrey Lijffijt, and Tijl De Bie. "ReCon: Reducing congestion in job recommendation using optimal transport." In Proceedings of the 17th ACM Conference on Recommender Systems, pp. 696-701. 2023.

  • Mashayekhi, Yoosof, Bo Kang, Jefrey Lijffijt, and Tijl De Bie. "GREASE: Graph Imbalance Reduction by Adding Sets of Edges." IEEE Transactions on Knowledge and Data Engineering 36, no. 4 (2023): 1611-1623.

  • Mashayekhi, Yoosof, Bo Kang, Jefrey Lijffijt, and Tijl De Bie. "Quantifying and reducing imbalance in networks." In RECSYS IN HR 2021, vol. 2967. CEUR, 2021.

  • Li, Nan, Bo Kang, Jefrey Lijffijt, and Tijl De Bie. "FEIR: Quantifying and reducing envy and inferiority for fair recommendation of limited resources." ACM Transactions on Intelligent Systems and Technology 15, no. 4 (2024): 1-24.

  • S. Nathan, G. Bied, E. Perennes, P. Caillou, B. Crepon, C. Gaillac, M. Sebag, JoLA: Job Landscape Aware Job Recommendation, forthcoming in the Proceedings of the 5th Workshop on Recommender Systems for Human Resources (RecSys-in-HR 2025), co-located with the 19th ACM Conference on Recommender Systems (RecSys 2025), Prague, Czech Republic, 22nd-26th September 2025

  • Mashayekhi, Yoosof, Nan Li, Bo Kang, Jefrey Lijffijt, and Tijl De Bie. "A challenge-based survey of e-recruitment recommendation systems." ACM Computing Surveys 56, no. 10 (2024): 1-33.

  • Victor Verreet, Lennert De Smet, Robin Manhaeve, Pieter Delobelle, and Jessa Bekker. "Inferring Missing CV Skills using PU Learning and Variational Inference." In International workshop on AI for Human Resources and Public Employment Services at ECML-PKDD 2023, Date: 2023/09/18-2023/09/22, Location: Turin, Italy. 2023.

Datasets

  • The public JobHop dataset of career trajectories.

Software & Tools

  • https://huggingface.co/datasets/aida-ugent/JobHop

  • https://aida-ugent.github.io/e-recruitment-recsys-challenges/

Team & Contact

Contact

UC Lead: Tijl De Bie, UGent, IDLab

AI Expert Lead: Tijl De Bie, UGent, IDLab

Use Case Lead(s)

Tijl De Bie, UGent, IDLab