Methods › General › Self-Supervised Learning
Self-Supervised Learning
The archive attaches this collection's text per method and the copies differ: 3 distinct texts across 43 of the 47 methods here. All are shown, most-carried first (a tie goes to the text carrying Papers with Code's collection boilerplate, then to the longer text); no vote is taken between them.
Text 1, carried by 40 of 47 methods:
Self-Supervised Learning refers to a category of methods where we learn representations in a self-supervised way (i.e without labels). These methods generally involve a pretext task that is solved to learn a good representation and a loss function to learn with. Below you can find a continuously updating list of self-supervised methods.
Text 2, carried by 2 of 47 methods:
Generative Models aim to model data generatively (rather than discriminatively), that is they aim to approximate the probability distribution of the data. Below you can find a continuously updating list of generative models for computer vision.
Text 3, carried by 1 of 47 methods:
The Graph Methods include neural network architectures for learning on graphs with prior structure information, popularly called as Graph Neural Networks (GNNs).
Recently, deep learning approaches are being extended to work on graph-structured data, giving rise to a series of graph neural networks addressing different challenges. Graph neural networks are particularly useful in applications where data are generated from non-Euclidean domains and represented as graphs with complex relationships.
Some tasks where GNNs are widely used include node classification, graph classification, link prediction, and much more.
In the taxonomy presented by Wu et al. (2019), graph neural networks can be divided into four categories: recurrent graph neural networks, convolutional graph neural networks, graph autoencoders, and spatial-temporal graph neural networks.
Image source: A Comprehensive Survey on Graph NeuralNetworks
Methods
All 47 methods in this collection, most-tagged first. Year is the archive's introduced_year; the archive stores 2000 when it has none, shown here as “–”. Papers counts distinct papers the archive tags with the method. Click a heading to sort.
| Inpainting | – | 998 |
| MAE Masked autoencoder | – | 722 |
| SimCLR | – | 240 |
| Colorization | – | 213 |
| DINO self-DIstillation with NO labels | – | 208 |
| MoCo Momentum Contrast | – | 148 |
| Jigsaw | – | 140 |
| BYOL Bootstrap Your Own Latent | – | 123 |
| Contrastive Predictive Coding | – | 113 |
| Barlow Twins | – | 71 |
| SwAV Swapping Assignments between Views | – | 55 |
| COLA | – | 32 |
| MoCo v2 | – | 30 |
| SEER | – | 19 |
| BiGAN Bidirectional GAN | – | 15 |
| NCL Neighborhood Contrastive Learning | – | 15 |
| CMCL Crossmodal Contrastive Learning | – | 12 |
| ReLIC | – | 12 |
| DeepCluster | – | 10 |
| SSDS Self-Supervised Deep Supervision | – | 10 |
| Dense Contrastive Learning | – | 9 |
| PIRL | – | 8 |
| M2D Masked Modeling Duo | – | 7 |
| MARLIN | – | 7 |
| Mirror-BERT | – | 6 |
| Contrastive Multiview Coding | – | 5 |
| NNCLR Nearest-Neighbor Contrastive Learning of Visual Representations | – | 5 |
| BigBiGAN | – | 4 |
| Graph Contrastive Coding | – | 4 |
| NPID | – | 4 |
| RotNet | – | 4 |
| CRISS | – | 3 |
| CVRL Contrastive Video Representation Learning | – | 3 |
| DeCLUTR | – | 3 |
| IMGEP Intrinsically Motivated Goal Exploration Processes | – | 2 |
| MoBY | – | 2 |
| PASE+ Problem Agnostic Speech Encoder + | – | 2 |
| ALP-GMM Absolute Learning Progress and Gaussian Mixture Models for Automatic Curriculum Learning | – | 1 |
| CPC v2 | – | 1 |
| ClusterFit | – | 1 |
| Internet Explorer | – | 1 |
| Magnification Prior Contrastive Similarity | – | 1 |
| NPID++ | – | 1 |
| ParamCrop | – | 1 |
| ReCo Regional Contrast | – | 1 |
| Trans-Encoder | – | 1 |
| VC R-CNN Visual Commonsense Region-based Convolutional Neural Network | – | 1 |