DeGMix: Efficient Multi-Task Dense Prediction with Deformable and Gating Mixer

TL;DR

DeGMix is an efficient multi-task dense prediction model that combines deformable CNNs and query-based Transformers with shared gating, achieving superior performance with lower computational cost.

Contribution

This work introduces DeGMix, a novel multi-task dense prediction framework integrating deformable CNNs and gating transformers for improved efficiency and accuracy.

Findings

DeGMix outperforms existing models on multiple dense prediction datasets.

DeGMix uses fewer GFLOPs than comparable Transformer and CNN models.

DeGMix achieves higher accuracy with less computational complexity.

Abstract

Convolution neural networks and Transformers have their own advantages and both have been widely used for dense prediction in multi-task learning (MTL). Existing studies typically employ either CNNs (effectively capture local spatial patterns) or Transformers (capturing long-range dependencies) independently, but integrating their strengths may yield more robust models. In this work, we present an efficient MTL model that combines the adaptive capabilities of deformable CNN and query-based Transformer with shared gating for MTL of dense prediction. This combination may offer a simple and efficient solution owing to its powerful and flexible task-specific learning and the advantages of lower cost, less complexity, and smaller parameters than traditional MTL methods. We introduce an efficient multi-task dense prediction with deformable and gating mixer (DeGMix). First, the deformable mixer encoder contains two types of operators: the channel-aware mixing operator leveraged to allow communication among different channels, and the spatial-aware deformable operator with deformable convolution applied to efficiently sample more informative spatial locations. Second, the task-aware gating transformer decoder is used to perform task-specific predictions, in which task interaction block integrated with self-attention is applied to capture task interaction features, and the task query block integrated with gating attention is leveraged to dynamically select the corresponding task-specific features. Furthermore, the results of the experiment demonstrate that the proposed DeGMix uses fewer GFLOPs and significantly outperforms current Transformer-based and CNN-based competitive models on a variety of metrics on three dense prediction datasets. Our code and models are available at https://github.com/yangyangxu0/DeMTG.