Convolutional Initialization for Data-Efficient Vision Transformers
- URL: http://arxiv.org/abs/2401.12511v1
- Date: Tue, 23 Jan 2024 06:03:16 GMT
- Title: Convolutional Initialization for Data-Efficient Vision Transformers
- Authors: Jianqiao Zheng, Xueqian Li, Simon Lucey
- Abstract summary: Training vision transformer networks on small datasets poses challenges.
CNNs can achieve state-of-the-art performance by leveraging their architectural inductive bias.
Our approach is motivated by the finding that random impulse filters can achieve almost comparable performance to learned filters in CNNs.
- Score: 38.63299194992718
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Training vision transformer networks on small datasets poses challenges. In
contrast, convolutional neural networks (CNNs) can achieve state-of-the-art
performance by leveraging their architectural inductive bias. In this paper, we
investigate whether this inductive bias can be reinterpreted as an
initialization bias within a vision transformer network. Our approach is
motivated by the finding that random impulse filters can achieve almost
comparable performance to learned filters in CNNs. We introduce a novel
initialization strategy for transformer networks that can achieve comparable
performance to CNNs on small datasets while preserving its architectural
flexibility.
Related papers
- DuoFormer: Leveraging Hierarchical Visual Representations by Local and Global Attention [1.5624421399300303]
We propose a novel hierarchical transformer model that adeptly integrates the feature extraction capabilities of Convolutional Neural Networks (CNNs) with the advanced representational potential of Vision Transformers (ViTs)
Addressing the lack of inductive biases and dependence on extensive training datasets in ViTs, our model employs a CNN backbone to generate hierarchical visual representations.
These representations are then adapted for transformer input through an innovative patch tokenization.
arXiv Detail & Related papers (2024-07-18T22:15:35Z) - Structured Initialization for Attention in Vision Transformers [34.374054040300805]
convolutional neural networks (CNNs) have an architectural inductive bias enabling them to perform well on small-scale problems.
We argue that the architectural bias inherent to CNNs can be reinterpreted as an initialization bias within ViT.
This insight is significant as it empowers ViTs to perform equally well on small-scale problems while maintaining their flexibility for large-scale applications.
arXiv Detail & Related papers (2024-04-01T14:34:47Z) - Self-Supervised Pre-Training for Table Structure Recognition Transformer [25.04573593082671]
We propose a self-supervised pre-training (SSP) method for table structure recognition transformers.
We discover that the performance gap between the linear projection transformer and the hybrid CNN-transformer can be mitigated by SSP of the visual encoder in the TSR model.
arXiv Detail & Related papers (2024-02-23T19:34:06Z) - NAR-Former V2: Rethinking Transformer for Universal Neural Network
Representation Learning [25.197394237526865]
We propose a modified Transformer-based universal neural network representation learning model NAR-Former V2.
Specifically, we take the network as a graph and design a straightforward tokenizer to encode the network into a sequence.
We incorporate the inductive representation learning capability of GNN into Transformer, enabling Transformer to generalize better when encountering unseen architecture.
arXiv Detail & Related papers (2023-06-19T09:11:04Z) - Pretraining Graph Neural Networks for few-shot Analog Circuit Modeling
and Design [68.1682448368636]
We present a supervised pretraining approach to learn circuit representations that can be adapted to new unseen topologies or unseen prediction tasks.
To cope with the variable topological structure of different circuits we describe each circuit as a graph and use graph neural networks (GNNs) to learn node embeddings.
We show that pretraining GNNs on prediction of output node voltages can encourage learning representations that can be adapted to new unseen topologies or prediction of new circuit level properties.
arXiv Detail & Related papers (2022-03-29T21:18:47Z) - Implicit Equivariance in Convolutional Networks [1.911678487931003]
Implicitly Equivariant Networks (IEN) induce equivariant in the different layers of a standard CNN model.
We show IEN outperforms the state-of-the-art rotation equivariant tracking method while providing faster inference speed.
arXiv Detail & Related papers (2021-11-28T14:44:17Z) - Can Vision Transformers Perform Convolution? [78.42076260340869]
We prove that a single ViT layer with image patches as the input can perform any convolution operation constructively.
We provide a lower bound on the number of heads for Vision Transformers to express CNNs.
arXiv Detail & Related papers (2021-11-02T03:30:17Z) - Container: Context Aggregation Network [83.12004501984043]
Recent finding shows that a simple based solution without any traditional convolutional or Transformer components can produce effective visual representations.
We present the model (CONText Ion NERtwok), a general-purpose building block for multi-head context aggregation.
In contrast to Transformer-based methods that do not scale well to downstream tasks that rely on larger input image resolutions, our efficient network, named modellight, can be employed in object detection and instance segmentation networks.
arXiv Detail & Related papers (2021-06-02T18:09:11Z) - On Robustness and Transferability of Convolutional Neural Networks [147.71743081671508]
Modern deep convolutional networks (CNNs) are often criticized for not generalizing under distributional shifts.
We study the interplay between out-of-distribution and transfer performance of modern image classification CNNs for the first time.
We find that increasing both the training set and model sizes significantly improve the distributional shift robustness.
arXiv Detail & Related papers (2020-07-16T18:39:04Z) - Curriculum By Smoothing [52.08553521577014]
Convolutional Neural Networks (CNNs) have shown impressive performance in computer vision tasks such as image classification, detection, and segmentation.
We propose an elegant curriculum based scheme that smoothes the feature embedding of a CNN using anti-aliasing or low-pass filters.
As the amount of information in the feature maps increases during training, the network is able to progressively learn better representations of the data.
arXiv Detail & Related papers (2020-03-03T07:27:44Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.