NiNformer: A Network in Network Transformer with Token Mixing Generated Gating Function
This work addresses efficiency issues in vision transformers for image classification, offering a novel method that is incremental in improving upon existing designs like MLP-Mixer.
The paper tackles the computational expense and data requirements of attention mechanisms in vision transformers by introducing a Network in Network Transformer block that replaces attention layers with a dynamic gating function generated through token mixing. The result is improved performance over baseline architectures on multiple image classification datasets, as shown through extensive experimentation.
The attention mechanism is the primary component of the transformer architecture; it has led to significant advancements in deep learning spanning many domains and covering multiple tasks. In computer vision, the attention mechanism was first incorporated in the Vision Transformer ViT, and then its usage has expanded into many tasks in the vision domain, such as classification, segmentation, object detection, and image generation. While the attention mechanism is very expressive and capable, it comes with the disadvantage of being computationally expensive and requiring datasets of considerable size for effective optimization. To address these shortcomings, many designs have been proposed in the literature to reduce the computational burden and alleviate the data size requirements. Examples of such attempts in the vision domain are the MLP-Mixer, the Conv-Mixer, the Perciver-IO, and many more attempts with different sets of advantages and disadvantages. This paper introduces a new computational block as an alternative to the standard ViT block. The newly proposed block reduces the computational requirements by replacing the normal attention layers with a Network in Network structure, therefore enhancing the static approach of the MLP-Mixer with a dynamic learning of element-wise gating function generated by a token mixing process. Extensive experimentation shows that the proposed design provides better performance than the baseline architectures on multiple datasets applied in the image classification task of the vision domain.