Self-Supervised Representation Learning: Introduction, advances, and challenges
Open accessThe DOI matches its Crossref record; 111 of the 125 references it cites were found in a registry or an index, and 1 are not checked yet.
Abstract
Self-supervised representation learning (SSRL) methods aim to provide powerful, deep feature learning without the requirement of large annotated data sets, thus alleviating the annotation bottleneck—one of the main barriers to the practical deployment of deep learning today. These techniques have advanced rapidly in recent years, with their efficacy approaching and sometimes surpassing fully supervised pretraining alternatives across a variety of data modalities, including image, video, sound, text, and graphs. This article introduces this vibrant area, including key concepts, the four main families of approaches and associated state-of-the-art techniques, and how self-supervised methods are applied to diverse modalities of data. We further discuss practical considerations including workflows, representation transferability, and computational cost. Finally, we survey major open challenges in the field, that provide fertile ground for future work.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source doesn’t support the claim as written.
In the citing paper
“The various self-supervised pretexts are divided into five broad families: transformation prediction, masked prediction, instance discrimination, clustering, and contrastive instance discrimination.”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [28]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Contrastive learning [56-58] is in fact a training procedure used in recent artificial intelligence models as a way to efficiently learn semantic representations.”
From How Adversarial REM Dreams May Facilitate Creativity, and Why We Become Aware of Them (2024) · cites this paper as [58]
The cited source says what the claim attributes to it.
In the citing paper
“For example, Ericsson et al. [28] provides a very comprehensive review of pretext tasks for multiple modalities, including image, video, text, audio, time series, and graph.”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [28]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 4×
“State of the art architectures usually start with CNN representation encoding hθ(·), with ResNet [1] being widely used, before appending task-specific”
InfluentialVaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404 · cited 3×
“long been used to model the context while recent works have moved to Transformer-based architectures with self-attention [2], which allow longer range connections to be made across words in a sentence, but require more data for training.”
Exploiting Generative AI to Scale up Intelligent Tutoring Systems
Jakubův, J., Chvalovský, K., Goertzel, Z., Kaliszyk, C., Olšák, M., Piotrowski, B., Schulz, S., Suda, M., & Urban, J. (2023) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.4230/lipics.itp.2023.19 · Cited by 76,658
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173 · cited 9×
“Self-supervised strategies have been leveraged successfully to improve sample efficiency of learning across a variety of modalities, from image [5]–[7], video [8], [9], speech [10], [11], text [12], [13], and graphs [14], [15].”
InfluentialImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 2×
“As in much of computer vision, ImageNet [35] is the most typical source data set for self-supervised pretraining [5]–[7], [40], consisting of 1.28 million training images across 1,000 object categories, with the most commonly used resolution at 224 × 224.”
InfluentialBERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325 · cited 9×
“Self-supervised strategies have been leveraged successfully to improve sample efficiency of learning across a variety of modalities, from image [5]–[7], video [8], [9], speech [10], [11], text [12], [13], and graphs [14], [15].”
InfluentialAn Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020) · arXiv · doi:10.48550/arxiv.2010.11929 · Cited by 47,472 · cited 1×
“One notable version is a vision transformer [58], which is increasingly being used by recent self-supervised methods on image data [43].”
The Elements of Statistical Learning
Hastie, T., Tibshirani, R., & Friedman, J. (2009) · Springer Series in Statistics · doi:10.1007/978-0-387-84858-7 · Cited by 25,970 · cited 2×
“These range from classic shallow approaches such as Gaussian mixtures [21] to deep methods such as variational autoencoders (VAEs) and generative adversarial networks (GANs) [18].”
Semi-supervised classification with graph convolutional networks
Kipf, T. N., & Welling, M. (2016) · arXiv · doi:10.48550/arxiv.1609.02907 · Cited by 24,853 · cited 1×
“Graph NNs (GNNs) [89] have advanced all these tasks significantly, especially where massive-labeled data are available.”
Distributed Representations of Words and Phrases and their Compositionality
Mikolov, T., Sutskever, I., Chen, K., Corrado, G., & Dean, J. (2013) · arXiv · doi:10.48550/arxiv.1310.4546 · Cited by 24,533 · cited 6×
“With data annotation being a major bottleneck, NLP was the first discipline to make major and successful use of self-supervision [22].”
Influential
Research works printed but not linked: 4
- 0, Youtube-8m A large and diverse labeled video dataset for video understanding research
- gokaslan, 2019, OpenWebText Corpus
- radford, 2019, Language Models are Unsupervised Multitask Learners
- goodfellow, 2016, Deep Learning
Cited by
Items 601–641 of 641
Mohammadi, M. R., Wang, H., Lin, R., Huang, W., Chen, Y., & Cui, Y. (2025) · International Conference on Intelligent Robotics and Control Engineering (IRCE) · doi:10.1109/irce66030.2025.11203182
“Representation learning has emerged as a powerful approach for extracting meaningful features from high-dimensional state spaces, demonstrating remarkable success in vision-based tasks [15].”
Georgios, P. (2024) · DSpace - NTUA (National Technical University of Athens) · doi:10.26240/heal.ntua.27199
Contrastive Learning for Fine-Grained Reading Detection
Islam, M. R., Vargo, A. W., Iwata, M., Iwamura, M., & Kise, K. (2025) · Information Engineering Express · doi:10.52731/iee.v11.i1.816
Edge AI and On-Device Machine Learning Optimization
Joel, B. (2022) · American International Journal of Computer Science and Technology · doi:10.63282/3117-5481/aijcst-v4i5p105
Research on positive and negative sample sampling strategies in contrastive learning
Liu, W. (2024) · Applied and Computational Engineering · doi:10.54254/2755-2721/101/20241021
Enhancing Automatic Polyp Segmentation Generalization Using Self-Supervised Learning
Monteiro, C. S. B. (2025) · Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT)
Débruitage et dématriçage concret de videos avec des réseaux de neurones
Dewil, V. (2022) · theses.fr (ABES) · doi:10.70675/dbf2e2b9zfec0z4a36z8866z18f80d92337e
Recommending activities for business process models
Sola, D. (2023) · MADOC (University of Mannheim)
Generating environments and pre-training agents for efficient reinforcement learning
Jelley, A. (2025) · ERA · doi:10.7488/era/6431
MOLE: MOdular Learning FramEwork via Mutual Information MaximizationPreprint
Li, T., & Pei, Y. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2308.07772
Wang, Y., Gao, Y., Zhang, G., Zhang, T., & Zhang, H. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-9911-7_15
Hu, Y. C. (2025) · Advances in Social Behavior Research · doi:10.54254/2753-7102/2025.25791
Torrents, J., Costa, T., & de Polavieja, G. G. (2025) · eLife · doi:10.7554/elife.107602.1.sa0
Video Anomaly Detection: A Systematic Taxonomy and Analysis of Deep Models
Xiao, L. (2025) · International Conference on Intelligent Computing (ICIC) · doi:10.65286/icic.v21i1.97344
Self-supervised learning for colloquial Finnish
Sõrmus, B. (2024) · Aaltodoc (Aalto University)
Mask-embedded transformer for English text recognition and correction
Sang, H. (2025) · International Journal of Information and Communication Technology · doi:10.1504/ijict.2025.149045
Adversarial contrastive with leveraging negative knowledge for point of interest sequence learning
Zhu, J., Luo, X., Wei, X., & Yao, X. (2026) · Neural Networks · doi:10.1016/j.neunet.2026.108599
CNN ‐Based Accent Similarity Detection Using Masked Spectrogram Reconstruction
Salifu, A., Nunoo‐Mensah, H., Tchao, E. T., Dede, A., Acheampong, F. A., Cobbinah, M., Adinkrah, J., Agbemenu, A. S., & Kponyo, J. J. (2026) · Engineering Reports · doi:10.1002/eng2.70665
“This trend persists in the latest literature: large self-supervised models achieve impressive cross-lingual transfer but require massive pre-training data and careful fine-tuning [20].”
Devika, A. K., Jose, B. R., Mathew, J., Thankachan, R. M., & Dayananda, P. (2026) · IEEE Access · doi:10.1109/access.2026.3675675
“By generating and completing pretext tasks, models can acquire significant feature representations from unlabeled data [54].”
Autoencoder Empowered EEG Data Classification: A Self-Supervised Learning Approach
Bakmeedeniya, A. H. M. T. C., Wijayakoon, W. B. M. S. C., & Ukgoda, U. (2024) · KDU International Research Conference (KDU IRC) · doi:10.64701/irc2024/345/8768
Wang, C., You, M., & Qin, F. (2026) · Journal of Environmental & Earth Sciences · doi:10.30564/jees.v8i4.13043
Self-Predictive Representation for Autonomous UAV Object-Goal NavigationPreprint
Ayala, A., Sui, D., Cruz, F., Torok, M., Deghat, M., & Fernandes, B. J. T. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2604.21130
“Self-predictive representation approaches arise from self-supervised learning, which optimizes an encoder model through a similarity measure between a given sample and a positive and negative subset [39].”
Zhang, Z., & Gong, F. (2025) · International Conference on Big Data, Artificial Intelligence and Risk Management (ICBAR) · doi:10.1145/3800227.3800231
“Representation learning [9] aims to automatically learn feature representations suitable for downstream tasks from raw data and has achieved remarkable success in domains involving continuous numerical data, such as image recognition [10], natural language processing [11], and speech recognition…”
Elliott, J. M. (2026) · Frontiers in Astronomy and Space Sciences · doi:10.3389/fspas.2026.1804102
CLIP-Based Approach for Zero-Shot Visual Recognition in Industrial Assembly Scenarios
Lisowska, S. K. (2026) · Data Engineering and Applications · doi:10.64972/dea.2026.v5i1.1624d:46-59
“On the other hand, these models have not been able to effectively address the open-set recognition problem in the complex visual and semantic environments of industrial assembly tasks [10].”
Wang, E. (2026) · Ingegneria Sismica · doi:10.65102/is20261064
EyeMVP: OCT-Informed Fundus Representation Learning via Paired CFP--OCT PretrainingPreprint
Deng, Z., Zhang, R., Zhang, Z., Gao, W., Li, Y., Wang, Q., Shao, L., Dong, J., Zeng, Z., 方立建, et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2606.15129
“Recent retinal foundation models have shown that self-supervised pretraining can substantially improve the transfer-ability of ophthalmic image representations [4]–[6].”
Anagnostaras, I.-D. (2026) · DSpace - NTUA (National Technical University of Athens) · doi:10.26240/heal.ntua.33093
Self-supervised extraction of perspective cues from hand-drawn sketches in environmental design
Han, N., & Zhang, L. (2026) · International Journal of Information and Communication Technology · doi:10.1504/ijict.2026.155636
Time-Aware Soft Actor–Critic via Observation Augmentation for Robust Continuous Control
Abdulsattar, S. S., Jaber, F. A., Al‐Balasmeh, H., & Mangoud, M. A. (2026) · International Conference on Frontiers of Engineering and Emerging Technologies (FET) · doi:10.1109/fet68771.2026.11601561
“Related approaches introduce auxiliary objectives, such as predicting future observations or rewards, learning latent dynamics, or self-supervised representation learning, with the goal of stabilizing training and improving sample efficiency under limited observability [12].”
Suneel, S., Swamy, B. V., Kashimalla, N., Swathi, G. L., Narang, M., & Upadhyay, S. (2026) · International Conference on Cognitive Computing and Networking Systems (ICC-CNS) · doi:10.1109/icc-cns70518.2026.11606022
“The video representation learning has experienced significant progress due to the recent advances in self-supervised learning that do not need the use of labeled data [3].”
Efficient ECG representation learning via intervention-based attribution alignment
Zhang, H., li, J., & Li, Y. (2026) · Physiological Measurement · doi:10.1088/1361-6579/ae9450
González, S., Sánchez, K., Saavedra, J. M., Pizarro, M., & Ghanem, B. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2608.24364
“It is now a common practice to use encoders trained under a self-supervised regimen because of their capability to learn from unlabeled data [5, 17, 1].”
Guarding the Open Seas: Autonomous Detection of AIS Signal Evasion Using Self -Supervised Learning
Sathish, K., Sangeetha, B., PavanKumar, C., Bai, D. S., & Spoorthi, K. S. (2026) · CRC Press eBooks · doi:10.1201/9781042005376-141
Zeng, C., Wang, F., & Wang, Z. (2026) · International Journal of Intelligent Systems · doi:10.1155/int/6478632
REDUKSI DATA BERLABEL PADA DETEKSI TUBERKULOSIS BERBASIS CITRA X-RAY MENGGUNAKAN FRAMEWORK SIMCLR
Majeri, M., Rachman, F. I., & Faisal, M. (2026) · Jurnal Informatika Progres · doi:10.56708/progres.v18i2.469
A self-supervised unified representation architecture for multi-source heterogeneous data
Tong, L., & Zhou, L. (2026) · International Journal of Reasoning-based Intelligent Systems · doi:10.1504/ijris.2026.157569
Collision detection in polishing robots based on self-supervised learning
Xing, Y., Zhang, F., Liu, D., & Mu, D. (2026) · Scientific Reports · doi:10.1038/s41598-026-78658-0
da Luz, G. P. C. P., Napoli, O. O., Delgado, J. V., Rocha, A. R., Boccato, L., & Borin, E. (2024) · Anais da Brazilian Conference on Intelligent Systems (BRACIS)
Exploring Data-Oriented Strategies for Training ECG Self-supervised Models
Silva, R. D. C., Medronha, A., & Silva, D. F. (2025) · Anais da Brazilian Conference on Intelligent Systems (BRACIS)
Tackling Low-Resource ECG Classification with Self-supervised Learning
Silva, R. D. C., & Silva, D. F. (2024) · Anais da Brazilian Conference on Intelligent Systems (BRACIS)
Metadata
- Authors
- Ericsson, L., Gouk, H., Loy, C. C., & Hospedales, T. M.
- Affiliations
- University of Edinburgh (GB)
- Nanyang Technological University (SG)
- Venue
- IEEE Signal Processing Magazine
- Published
- 2022-05
- Volume
- 39
- Issue
- 3
- Pages
- 42-62
- ISSN
- 1053-5888 (ISSN-L)
- 1558-0792
- OpenAlex
- W3207924272
- Cited by
- 433
- References
- 125
- Citations by year
- 2026: 89 · 2025: 105 · 2024: 132 · 2023: 86 · 2022: 20
- FWCI
- 28.22
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Signal Processing Magazine (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- Edinburgh Research Explorer (University of Edinburgh) (submittedVersion, cc-by-nd)
- DR-NTU (Nanyang Technological University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Engineering and Physical Sciences Research Council (EP/S000631/1, EP/L016427/1, EP/R026173/1)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Physical Sciences › Computer Science › Artificial Intelligence › Advanced Graph Neural Networks
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Algorithms
- Keywords
- self-supervised representation learning, pre-training, deep feature learning, image representation, video representation, sound representation, text representation, graph representation, computational cost, open problems
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.