The Empirical Impact of Reducing Symmetries on the Performance of Deep Ensembles and MoE
This work addresses the practical impact of symmetry reduction for ensemble methods in machine learning, though it appears incremental as it builds on prior studies of linear mode connectivity.
The paper investigated how reducing symmetries in neural networks affects deep ensembles and Mixture of Experts (MoE) across five datasets, finding that asymmetric networks significantly improve deep ensemble performance as size increases, but results for MoE and a new Mixture of Interpolated Experts (MoIE) were inconclusive.
Recent studies have shown that reducing symmetries in neural networks enhances linear mode connectivity between networks without requiring parameter space alignment, leading to improved performance in linearly interpolated neural networks. However, in practical applications, neural network interpolation is rarely used; instead, ensembles of networks are more common. In this paper, we empirically investigate the impact of reducing symmetries on the performance of deep ensembles and Mixture of Experts (MoE) across five datasets. Additionally, to explore deeper linear mode connectivity, we introduce the Mixture of Interpolated Experts (MoIE). Our results show that deep ensembles built on asymmetric neural networks achieve significantly better performance as ensemble size increases compared to their symmetric counterparts. In contrast, our experiments do not provide conclusive evidence on whether reducing symmetries affects both MoE and MoIE architectures.