Self-Supervised Image Classification
태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.
- 하위 태스크
- 0
- 평가표
- 3
- 보존된 평가 기록
- 208
This is the task of image classification using representations learnt with self-supervised learning. Self-supervised methods generally involve a pretext task that is solved to learn a good representation and a loss function to learn with. One example of a loss function is an autoencoder based loss where the goal is reconstruction of an image pixel-by-pixel. A more popular recent example is a contrastive loss, which measure the similarity of sample pairs in a representation space, and where there can be a varying target instead of a fixed target to reconstruct (as in the case of autoencoders).
A common evaluation protocol is to train a linear classifier on top of (frozen) representations learnt by self-supervised methods. The leaderboards for the linear evaluation protocol can be found below. In practice, it is more common to fine-tune features on a downstream task. An alternative evaluation protocol therefore uses semi-supervised learning and finetunes on a % of the labels. The leaderboards for the finetuning protocol can be accessed here.
You may want to read some blog posts before reading the papers and checking the leaderboards:
- Contrastive Self-Supervised Learning - Ankesh Anand
- The Illustrated Self-Supervised Learning - Amit Chaudhary
- Self-supervised learning and computer vision - Jeremy Howard
- Self-Supervised Representation Learning - Lilian Weng
There is also Yann LeCun's talk at AAAI-20 which you can watch here (35:00+).
<span style="color:grey; opacity: 0.6">( Image credit: A Simple Framework for Contrastive Learning of Visual Representations )</span>
논문
이 태스크와 하위 분야의 관련 논문입니다. 실제 성능 비교는 평가 결과에서 확인하세요.
Weak Augmentation Guided Relational Self-Supervised Learning
Vision Models Are More Robust And Fair When Pretrained On Uncurated Images Without Supervision
Context Autoencoder for Self-Supervised Representation Learning
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework
When Do Flat Minima Optimizers Work?
Pushing the limits of self-supervised ResNets: Can we outperform supervised learning without labels on ImageNet?
Max-Margin Contrastive Learning
Masked Feature Prediction for Self-Supervised Visual Pre-Training
Similarity Contrastive Estimation for Self-Supervised Soft Contrastive Learning
PeCo: Perceptual Codebook for BERT Pre-training of Vision Transformers
SimMIM: A Simple Framework for Masked Image Modeling
iBOT: Image BERT Pre-Training with Online Tokenizer
Masked Autoencoders Are Scalable Vision Learners
Self-Supervised Learning by Estimating Twin Class Distributions
Weakly Supervised Contrastive Learning
Compressive Visual Representations
ReSSL: Relational Self-Supervised Learning with Weak Augmentation
XCiT: Cross-Covariance Image Transformers
Efficient Self-supervised Vision Transformers for Representation Learning
BEiT: BERT Pre-Training of Image Transformers
Unsupervised Visual Representation Learning by Online Constrained K-Means
Large-Scale Unsupervised Person Re-Identification with Contrastive Learning
Divide and Contrast: Self-supervised Learning from Uncurated Data
VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learning
Self-Supervised Learning with Swin Transformers
ResMLP: Feedforward networks for image classification with data-efficient training
Emerging Properties in Self-Supervised Vision Transformers
With a Little Help from My Friends: Nearest-Neighbor Contrastive Learning of Visual Representations
Mutual Contrastive Learning for Visual Representation Learning
Solving Inefficiency of Self-supervised Representation Learning
Contrastive Learning with Stronger Augmentations
An Empirical Study of Training Self-Supervised Vision Transformers
Self-Supervised Classification Network
Barlow Twins: Self-Supervised Learning via Redundancy Reduction
Self-supervised Pretraining of Visual Features in the Wild
출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.