Multi-Domain Learning with Global Expert Mapping

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Shamsolmoali, Pourya, Zareapoor, Masoumeh, Zhou, Huiyu, Mendez, Oscar, Tao, Dacheng, Li, Xuelong
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866918458910834688
author Shamsolmoali, Pourya
Zareapoor, Masoumeh
Zhou, Huiyu
Mendez, Oscar
Tao, Dacheng
Li, Xuelong
author_facet Shamsolmoali, Pourya
Zareapoor, Masoumeh
Zhou, Huiyu
Mendez, Oscar
Tao, Dacheng
Li, Xuelong
contents Human perception generalizes well across different domains, but most vision models struggle beyond their training data. This gap motivates multi-dataset learning, where a single model is trained on diverse datasets to improve robustness under domain shifts. However, unified training remains challenging due to inconsistencies in data distributions and label semantics. Mixture-of-Experts (MoE) models provide a scalable solution by routing inputs to specialized subnetworks (experts). Yet, existing MoEs often fail to specialize effectively, as their load-balancing mechanisms enforce uniform input distribution across experts. This fairness conflicts with domain-aware routing, causing experts to learn redundant representations, and reducing performance especially on rare or out-of-distribution domains. We propose GEM (Global Expert Mapping), a planner-compiler framework that replaces the learned router with a global scheduler. Our planner, based on linear programming relaxation, computes a fractional assignment of datasets to experts, while the compiler applies hierarchical rounding to convert this soft plan into a deterministic, capacity-aware mapping. Unlike prior MoEs, GEM avoids balancing loss, resolves the conflict between fairness and specialization, and produces interpretable routing. Experiments show that GEM-DINO achieves state-of-the-art performance on the UODB benchmark, with notable gains on underrepresented datasets and solves task interference in few-shot adaptation scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2604_18842
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Multi-Domain Learning with Global Expert Mapping
Shamsolmoali, Pourya
Zareapoor, Masoumeh
Zhou, Huiyu
Mendez, Oscar
Tao, Dacheng
Li, Xuelong
Computer Vision and Pattern Recognition
Human perception generalizes well across different domains, but most vision models struggle beyond their training data. This gap motivates multi-dataset learning, where a single model is trained on diverse datasets to improve robustness under domain shifts. However, unified training remains challenging due to inconsistencies in data distributions and label semantics. Mixture-of-Experts (MoE) models provide a scalable solution by routing inputs to specialized subnetworks (experts). Yet, existing MoEs often fail to specialize effectively, as their load-balancing mechanisms enforce uniform input distribution across experts. This fairness conflicts with domain-aware routing, causing experts to learn redundant representations, and reducing performance especially on rare or out-of-distribution domains. We propose GEM (Global Expert Mapping), a planner-compiler framework that replaces the learned router with a global scheduler. Our planner, based on linear programming relaxation, computes a fractional assignment of datasets to experts, while the compiler applies hierarchical rounding to convert this soft plan into a deterministic, capacity-aware mapping. Unlike prior MoEs, GEM avoids balancing loss, resolves the conflict between fairness and specialization, and produces interpretable routing. Experiments show that GEM-DINO achieves state-of-the-art performance on the UODB benchmark, with notable gains on underrepresented datasets and solves task interference in few-shot adaptation scenarios.
title Multi-Domain Learning with Global Expert Mapping
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.18842