Double-Stage Feature-Level Clustering-Based Mixture of Experts Framework
This work addresses a specific bottleneck in MoE models for image classification, offering an incremental improvement over existing methods.
The paper tackles the problem of noise and outliers affecting Mixture-of-Experts (MoE) models in image classification by introducing the DFCP-MoE framework, which uses feature-level clustering and pseudo-labeling to improve expert specialization and achieves competitive inference results on three benchmark datasets.
The Mixture-of-Experts (MoE) model has succeeded in deep learning (DL). However, its complex architecture and advantages over dense models in image classification remain unclear. In previous studies, MoE performance has often been affected by noise and outliers in the input space. Some approaches incorporate input clustering for training MoE models, but most clustering algorithms lack access to labeled data, limiting their effectiveness. This paper introduces the Double-stage Feature-level Clustering and Pseudo-labeling-based Mixture of Experts (DFCP-MoE) framework, which consists of input feature extraction, feature-level clustering, and a computationally efficient pseudo-labeling strategy. This approach reduces the impact of noise and outliers while leveraging a small subset of labeled data to label a large portion of unlabeled inputs. We propose a conditional end-to-end joint training method that improves expert specialization by training the MoE model on well-labeled, clustered inputs. Unlike traditional MoE and dense models, the DFCP-MoE framework effectively captures input space diversity, leading to competitive inference results. We validate our approach on three benchmark datasets for multi-class classification tasks.