GEM replaces learned routers in MoE models with a global scheduler based on linear programming relaxation and hierarchical rounding, achieving SOTA on the UODB multi-domain benchmark with gains on rare domains.
Harder tasks need more experts: Dynamic routing in moe models,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Multi-Domain Learning with Global Expert Mapping
GEM replaces learned routers in MoE models with a global scheduler based on linear programming relaxation and hierarchical rounding, achieving SOTA on the UODB multi-domain benchmark with gains on rare domains.