Task-customized Masked AutoEncoder via Mixture of Cluster-conditional Experts

TL;DR

This paper introduces MoCE, a novel pre-training paradigm for Masked Autoencoders that creates task-specific models by leveraging cluster-conditional experts, improving transferability and performance across diverse downstream tasks.

Contribution

MoCE is the first to train a single MAE-based model with cluster-conditional experts, enabling customized pre-training for different downstream tasks.

Findings

MoCE outperforms vanilla MAE by 2.45% on average across 11 tasks.

Achieves state-of-the-art results on detection and segmentation.

Effective in reducing negative transfer from irrelevant pre-training data.

Abstract

Masked Autoencoder~(MAE) is a prevailing self-supervised learning method that achieves promising results in model pre-training. However, when the various downstream tasks have data distributions different from the pre-training data, the semantically irrelevant pre-training information might result in negative transfer, impeding MAE's scalability. To address this issue, we propose a novel MAE-based pre-training paradigm, Mixture of Cluster-conditional Experts (MoCE), which can be trained once but provides customized pre-training models for diverse downstream tasks. Different from the mixture of experts (MoE), our MoCE trains each expert only with semantically relevant images by using cluster-conditional gates. Thus, each downstream task can be allocated to its customized model pre-trained with data most similar to the downstream data. Experiments on a collection of 11 downstream tasks show that MoCE outperforms the vanilla MAE by 2.45\% on average. It also obtains new state-of-the-art self-supervised learning results on detection and segmentation.