paper-with-me

Papers

Design and Behavior of Sparse Mixture-of-Experts Layers in CNN-based Semantic Segmentation

2026-04-15 · Svetlana Pavlitska, Haixi Fan, Konstantin Ditschuneit, J. Marius Zöllner arxiv

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/.

📄 PDF Abstract BibTeX arXiv:2604.13761

Code (0)

등록된 구현이 없습니다.

Tasks

Semantic Segmentation

Similar Papers 제목 키워드 기반

MoEC: Mixture of Expert Clusters

2022-07-19 · Yuan Xie, Shaohan Huang, Tianyu Chen, Furu Wei

Sparsely Mixture of Experts (MoE) has received great interest due to its promising scaling capability with affordable computational overhead. MoE converts dense layers into sparse experts, and utilizes a gated routing ne…

Machine TranslationMixture-of-ExpertsNatural Language Understanding

Mixture of A Million Experts

2024-07-04 · Xu Owen He

The feedforward (FFW) layers in standard transformer architectures incur a linear increase in computational costs and activation memory as the hidden layer width grows. Sparse mixture-of-experts (MoE) architectures have …

Computational EfficiencyLanguage ModelingLanguage ModellingMixture-of-Experts+1

Secret mixtures of experts inside your LLM

2025-12-20 · Enric Boix-Adsera arxiv

Despite being one of the earliest neural network layers, the Multilayer Perceptron (MLP) is arguably one of the least understood parts of the transformer architecture due to its dense computation and lack of easy visuali…

Sparsely-gated Mixture-of-Expert Layers for CNN Interpretability

2022-04-22 · Svetlana Pavlitska, Christian Hubschneider, Lukas Struppek, J. Marius Zöllner

Sparsely-gated Mixture of Expert (MoE) layers have been recently successfully applied for scaling large transformers, especially for language modeling tasks. An intriguing side effect of sparse MoE layers is that they co…

image-classificationImage ClassificationLanguage ModelingLanguage Modelling+2

Robust Experts: the Effect of Adversarial Training on CNNs with Sparse Mixture-of-Experts Layers

2025-09-05 · Svetlana Pavlitska, Haixi Fan, Konstantin Ditschuneit, J. Marius Zöllner arxiv

Robustifying convolutional neural networks (CNNs) against adversarial attacks remains challenging and often requires resource-intensive countermeasures. We explore the use of sparse mixture-of-experts (MoE) layers to imp…