arXiv · 2604.13761
Design and Behavior of Sparse Mixture-of-Experts Layers in CNN-based Semantic Segmentation
Abstract
Sparse mixture-of-experts (MoE) layers have been shown to substantially increase model capacity without a proportional increase in computational cost and are widely used in transformer architectures, where they typically replace feed-forward network blocks. In contrast, integrating sparse MoE layers into convolutional neural networks (CNNs) remains inconsistent, with most prior work focusing on fine-grained MoEs operating at the filter or channel levels. In this work, we investigate a coarser, patch-wise formulation of sparse MoE layers for semantic segmentation, where local regions are routed to a small subset of convolutional experts. Through experiments on the Cityscapes and BDD100K datasets using encoder-decoder and backbone-based CNNs, we conduct a design analysis to assess how architectural choices affect routing dynamics and expert specialization. Our results demonstrate consistent, architecture-dependent improvements (up to +3.9 mIoU) with little computational overhead, while revealing strong design sensitivity. Our work provides empirical insights into the design and internal dynamics of sparse MoE layers in CNN-based dense prediction. Our code is available at https://github.com/KASTEL-MobilityLab/moe-layers/.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Svetlana Pavlitska, Haixi Fan, Konstantin Ditschuneit, J. Marius Zöllner. 2026-04-15. Design and Behavior of Sparse Mixture-of-Experts Layers in CNN-based Semantic Segmentation. https://arxiv.org/abs/2604.13761
Cite the original work for its findings. Save a collection to share your selection of sources.