Learned Cross-Task Relationships in Multi-Task Models

Abstract:We propose a framework that learns cross-task relationships in multi-task models by approximating the joint distribution of task labels through targeted pairwise relationships. This approach improves performance via transfer learning and enhances information extraction without the intractable complexity of modeling the full joint space. Although our framework applies to any multi-task system, we demonstrate its efficacy within YouTube's production recommendation systems. Experiments across the Notifications, Homepage, and Watch Next surfaces show improvements in both accuracy and user satisfaction metrics. Finally, we propose a workflow template to facilitate broader future implementation.

Accepted to the 20th ACM Conference on Recommender Systems (RecSys 2026). 10 pages, 6 figures

Artificial Intelligence (cs.AI)

arXiv:2609.28776 [cs.AI]

(or arXiv:2609.28776v1 [cs.AI] for this version)

https://doi.org/10.48550/arXiv.2609.28776 Focus to learn more arXiv-issued DOI via DataCite (pending registration)

https://doi.org/10.1145/3773078.3831878 Focus to learn more DOI(s) linking to related resources

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.