Contrastive Representation Learning: A Framework and Review
Open accessThe DOI matches its Crossref record; 147 of the 162 references it cites were found in a registry or an index, and 15 are not checked yet.
Abstract
Contrastive Learning has recently received interest due to its success in self-supervised representation learning in the computer vision domain. However, the origins of Contrastive Learning date as far back as the 1990s and its development has spanned across many fields and domains including Metric Learning and natural language processing. In this paper, we provide a comprehensive literature review and we propose a general Contrastive Representation Learning framework that simplifies and unifies many different contrastive learning methods. We also provide a taxonomy for each of the components of contrastive learning in order to summarise it and distinguish it from other forms of machine learning. We then discuss the inductive biases which are present in any contrastive learning system and we analyse our framework under different views from various sub-fields of Machine Learning. Examples of how contrastive learning has been applied in computer vision, natural language processing, audio processing, and others, as well as in Reinforcement Learning are also presented. Finally, we discuss the challenges and some of the most promising future research directions ahead.
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 source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“According to the suggestions of Le-Khac et al. [29], the main methods can be divided into the following categories: multisensory signals, data augmentation, localglobal consistency, and temporal consistency.”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [29]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
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 [56]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“As summarized in [29], commonly used contrastive loss functions generally include scoring functions (cosine similarity), energy-based margin functions (pair loss and triplet loss), probabilistic NCE-based functions, and mutual information based functions.”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [29]
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 2×
“In SimCLRv2 [17], empirical experiments show that the output of the second layer of a 3-layer MLP projection head is a better representation for transfer learning than the output of the base ResNet [42] encoder.”
BackgroundMethodVaswani, 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×
“Recently, the Bidirectional Encoder Representation from Transformer (BERT) [24] model learns bidirectional word representations using the Transformer architecture’s decoder [103] and demonstrated great performance for transfer learning in multiple downstream tasks.”
MethodHochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 1×
“Even though the BERT [24] text encoder and the Faster RCNN [87] object detection model are trained separately and not updated by the contrastive loss, the LSTM cells [48] and a linear project head can still map between words and object representations.”
MethodExploiting 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
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 73,472
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 5×
“The discretised speech representation can then be used directly as a representation for other models that expect discrete input such as BERT [24].”
MethodInfluentialFaster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2015) · arXiv · doi:10.48550/arxiv.1506.01497 · Cited by 70,812
ImageNet: 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
BERT: 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 5×
“The discretised speech representation can then be used directly as a representation for other models that expect discrete input such as BERT [24].”
MethodInfluentialEfficient Estimation of Word Representations in Vector Space
Mikolov, T., Chen, K., Corrado, G., & Dean, J. (2013) · arXiv · doi:10.48550/arxiv.1301.3781 · Cited by 26,568
Research works printed but not linked: 1
- stratos, 2019, Noise contrastive estimation
Cited by
Items 701–800 of 948
Wang, J., & Luo, L. (2026) · Alexandria Engineering Journal · doi:10.1016/j.aej.2026.02.016 · Cited by 1
Integrating Higher-Level Semantics into Robust Biomedical Name Representations
Fivez, P., Šuster, S., & Daelemans, W. M. P. (2021) · Cited by 1
Scaling Deep Contrastive Learning Batch Size with Almost Constant Peak Memory Usage.Preprint
Gao, L., & Zhang, Y. (2021) · arXiv (Cornell University) · Cited by 1
Generative Adversarial Learning via Kernel Density DiscriminationPreprint
Lemkhenter, A., Bielski, A., Sari, A. E., & Favaro, P. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.06197 · Cited by 1
Manifold Characteristics That Predict Downstream Task PerformancePreprint
van der Merwe, R., Newman, G., & Barnard, E. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2205.07477 · Cited by 1
“…techniques involve contrastive learning where representations from images are directly contrasted against one another such that the network learns to encode similar images to similar representations (Schroff et al., 2015; Oord et al., 2018; Chen et al., 2020; He et al., 2020; Le-Khac et al., 2020).”
Semi-Supervised Contrastive Outlier Removal for Pseudo Expectation Maximization (Scope)Preprint
Menon, S., & Chapman, D. (2022) · SSRN Electronic Journal · doi:10.2139/ssrn.4146348 · Cited by 1
PaDPaF: Partial Disentanglement with Partially-Federated GANsPreprint
Almansoori, A. J., Horváth, S., & Takáč, M. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2212.03836 · Cited by 3
“One well-known method is to use a contrastive objective [27], which simply pulls similar data together and pushes different data away.”
MethodZhang, Q. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2310.13710 · Cited by 1
“Reinforcement learning [11], supervised learning [12], and unsupervised learning [13] paradigms are incorporated to facilitate experiential learning and knowledge acquisition.”
BackgroundSellner, M. S., Mahmoud, A. H., & Lill, M. A. (2023) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2023.11.17.567506 · Cited by 1
“A common approach is contrastive representation learning, in which similar samples are trained to be close together in embedding space, while dissimilar samples should be farther apart [23].”
BackgroundA Critical Analysis of Unsupervised Learning in the Context of Raven's Progressive MatricesPreprint
Omari, R. M., McKay, B., Gedeon, T., & Taylor, K. L. (2024) · doi:10.36227/techrxiv.171198231.11568718/v1 · Cited by 1
How Adversarial Rem-Dreams May Facilitate Creativity, and Why We Become Aware of ThemPreprint
Deperrois, N., Petrovici, M. A., Jordan, J., Huber, L. S., & Senn, W. (2024) · Preprints.org · doi:10.20944/preprints202403.0684.v2 · Cited by 1
A Recommendation Method Based on Fusion of Graph Neural Network Single-Layer Mixing Negative SamplesPreprint
Cai, B., Zhou, K., Cheng, J., Tang, X., Hu, Y., Liu, M., & Miao, Q. (2024) · Research Square · doi:10.21203/rs.3.rs-3928759/v1 · Cited by 1
SimSiam Naming Game: A Unified Approach for Emergent Communication and Representation LearningPreprint
Hoang, N. L., Taniguchi, T., Tianwei, F., Taniguchi, A., & Nagano, M. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2410.21803 · Cited by 2
“One important approach in SSL is contrastive learning, which focuses on learning by comparing different augmented views of the same data point (Le-Khac et al., 2020).”
Analytical Software Engineering: A Novel Design Paradigm for Problem Formulation and OptimizationPreprint
Houichime, T. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5423734 · Cited by 1
A Joint Learning Approach to Hardware Caching and PrefetchingPreprint
Yuan, S., Saxena, D., Chen, J. K., Sharma, N., & Akella, A. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.10862 · Cited by 1
“The second approach relies on contrastive learning [1, 4, 8, 13, 14], where the encoder is trained by contrasting feature pairs that are temporally co-occurring as positive pairs against feature pairs that do not occur together as negative pairs.”
Rieger, W. J., Bodén, M., Arnold, F. H., & Mora, A. (2026) · eLife · doi:10.7554/elife.108186.2 · Cited by 1
HEDeST: An Integrative Approach to Enhance Spatial Transcriptomic Deconvolution with HistologyPreprint
Gortana, L., Chadoutaud, L., Bourgade, R., Barillot, E., & Walter, T. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.01.06.697922 · Cited by 1
“In order to obtain a robust representation of cells, we use a contrastive learning approach [57].”
Martino, L. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2604.08116 · Cited by 1
“Owing to its effectiveness and flexibility, NCE has been widely studied and applied in a variety of settings [17, 18, 19].”
Alshami, A., & Choi, J. (2025) · CIB Conferences · doi:10.7771/3067-4883.2103 · Cited by 1
FROST: Faster and more Robust One-shot Semi-supervised TrainingPreprint
Liu, H. E., & Smith, L. N. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.09471
“We investigated how well CRL works in replacing consistency regularization in the semi-supervised domain and discovered that CRL is a poor substitute for consistency regularization.”
BackgroundMethodInfluentialJornada de Atualização em Informática 2024
Cappelli, C., Alchieri, E. A. P., Campos, J. N., Mendonça, R. D., Fontinele, A., de Carvalho, L. H. S., Oliveira, I. R., Cardoso, Í. W. F., Coelho, R., Freitas, A. E. S., et al. (2024) · doi:10.5753/sbc.14597.4
Improving sound retrieval in large collaborative collections
Favory, X. (2021) · Repositori digital de la UPF (Universitat Pompeu Fabra) · Cited by 2
Sharp Learning Bounds for Contrastive Unsupervised Representation Learning.Preprint
Bao, H., Nagano, Y., & Nozawa, K. (2021) · arXiv (Cornell University)
Unsupervised Semantic Segmentation with Contrastive Translation Coding
Chen, R., Sun, H., & Wang, L. (2022) · Communications in computer and information science · doi:10.1007/978-981-16-9247-5_2
Asymptotics of $\ell_2$ Regularized Network EmbeddingsPreprint
Davison, A. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2201.01689
Barak, T., & Loewenstein, Y. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2205.01924
Pardo, J. G. G. (2022) · RiuNet (Universitat Politècnica de València) · doi:10.4995/thesis/10251/182400 · Cited by 1
Kim, T. H., Jang, W. S., Heo, S. C., & Park, Y. R. (2022) · doi:10.2196/preprints.41917
Boosting Self-supervised Video-based Human Action Recognition Through Knowledge Distillation
Camarena, F., González-Mendoza, M., Chang, L., & Hernández-Gress, N. (2022) · LatinX in AI at Neural Information Processing Systems Conference (LXAI) · doi:10.52591/lxai202211286
“On the other hand, contrastive learning [16] identifies what aspects of a video sample make it different from other samples.”
Backgroundcurr2vib: Modality Embedding Translation for Broken-Rotor Bar Detection
Berenji, A., Taghiyarrenani, Z., & Nowaczyk, S. (2023) · Communications in computer and information science · doi:10.1007/978-3-031-23633-4_28
Contrastive Learning for Topic-Dependent Image Ranking
Ko, J., Jeong, J., & Kim, K. (2023) · Lecture notes in electrical engineering · doi:10.1007/978-3-031-22192-7_5
Mengistu, M. T., Alemu, G., Chevaillier, P., & Loor, P. D. (2022) · International Conference on Machine Learning and Applications · doi:10.1109/icmla55696.2022.00273
“Contrastive representation learning is a learning by comparing, it uses mutual information (MI) estimation for making the similarity/dissimilarity comparisons between samples [26].”
Facial Expression Recognition and the Application of Supervised Contrastive Learning
Yi, C. (2022) · International Symposium on Advances in Informatics, Electronics and Education (ISAIEE) · doi:10.1109/isaiee57420.2022.00126
“A thorough survey of the literature is provided by Le-Khac et al[30], and they also suggest a basic framework for contrastive representation learning that simplifies and unifies a wide range of contrastive learning techniques.”
BackgroundLearning Representations of Satellite Images with Evaluations on Synoptic Weather EventsPreprint
Yo, T., Su, S., Wu, C., Chen, W., Chu, J., Chang, C., & Kuo, H. (2023) · doi:10.22541/essoar.168394729.95734739/v1
“Moreover, researchers have developed core principles of good representations, i.e., distributed , disentangled , and abstract and invariant (Bengio et al., 2013; Le-Khac et al., 2020).”
Gao, M., Tian, S., Yu, L., zhai, Y., & Long, J. (2023) · Research Square · doi:10.21203/rs.3.rs-3045297/v1
Embedding Clustering using Lightweight Contrastive Learning For Cross-Modal Classification
Amendola, J., Cenkeramaddi, L. R., & Jha, A. (2023) · International Conference on Computer and Communication Systems (ICCCS) · doi:10.1109/icccs57501.2023.10150770
“The mapping relies on the contrastive learning framework [20] to obtain encoders that can properly ”push together” (and ”pull apart”) embedding vectors according to the semantic similarity (and dissimilarities) of their respective inputs.”
BackgroundSpeaker recognition system of flexible throat microphone using contrastive learning
Zheng, W., Chen, Z., Li, Y., Jiang, X., & Cao, X. (2023) · IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid) · doi:10.1109/ccgrid57682.2023.00065 · Cited by 1
“Compared with other self-supervised representation learning, contrastive learning is much simpler and more efficient since it does not have to derive a pseudo-label from a pretext task or modify a model architecture during learning [26].”
BackgroundUnsupervised Learning of Invariance TransformationsPreprint
Vučković, A., Stock, B., Hopp, A. V., Winkel, M., & Linde, H. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2307.12937
Learning Invariant Representation Improves Robustness for MRC Models
Hai, Y., Liang, W., Meng, H., Liu, T., & Wang, H. (2022) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2022.findings-emnlp.241
“To be capable of handling more general attacks rather than just a certain type attack without laborious iterative schedule, we propose SCQA to addresses the above robustness issue by learning invariant representations of similar examples inspired by Le-Khac et al. (2020).”
MethodInfluentialUniversal Hidden Monotonic Trend Estimation with Contrastive Learning
Pineau, E., Razakarivony, S., Gonzalez, M., & Schrapffer, A. (2023) · Lecture notes in networks and systems · doi:10.1007/978-3-031-37963-5_36
MCRLe: Multi-Modal Contrastive Representation Learning For Stroke Onset Time Diagnosis
Liao, W., Jiang, P., Lv, Y., Xue, Y., Chen, Z., & Li, X. (2023) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi53787.2023.10230570
“1), a supervised model with contrastive learning constraint [7] that trains DNN on multi-modal medical images and boosts performance on stroke onset time diagnosis, for the following reasons: (1) Contrastive learning measure the distance between two representations using Noise-contrastive estimation (NCE) [8], which t…”
BackgroundMandache, D., la Guillaume, É. B. À., Olivo‐Marín, J., & Meas‐Yedid, V. (2023) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi53787.2023.10230713
“This endeavor falls under the umbrella of representation learning and is achieved via siamese convolutional neural networks [3] trained under a contrastive learning [4] paradigm.”
BackgroundGeneralizing Similarity in Noisy Setups: The DIBS Phenomenon
Fonseca, N., & Guidetti, V. (2023) · Frontiers in artificial intelligence and applications · doi:10.3233/faia230339
“Despite its success, [39, 32, 29] show that contrastive learning usually requires huge datasets and considerable use of data augmentation techniques.”
BackgroundSelf-supervised learning in different domains
Kamaleeva, A. M., & Bikmullina, I. I. (2023) · AIP conference proceedings · doi:10.1063/5.0175305
Contrastive Learning for Multivariate Time Series Classification: an Early Fusion Approach
Cagliero, L., Buccafusco, S., & Vaccarino, F. (2023) · IEEE International Conference on Application of Information and Communication Technologies (AICT) · doi:10.1109/aict59525.2023.10313147
“The purpose of CL is to first apply data augmentation to define positive and negative pairs of samples and then generate a latent representation where positive pairs of samples are placed relatively closed to each other, whereas negative ones are quite distant [14].”
BackgroundFeng, X., Li, C., & Zhang, S. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4665286 · Cited by 1
Prosody-Based Vocal Emotional Alignment in Infant-Caregiver Interaction
Li, M., Li, J., Zhang, D., & Nagai, Y. (2023) · IEEE International Conference on Development and Learning (ICDL) · doi:10.1109/icdl55364.2023.10364533
“Among these models, those based on contrastive learning algorithms [20] have emerged as prominent tools for directly comparing the nuances of vocal emotions.”
BackgroundMethodDeep Representation Learning of User Preferences for OpinionclassificationPreprint
Gunathilaka, T. M. A. U., Li, Y., Zhang, J., & Kelly, W. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4676404
Encoding Binary Events from Continuous Time Series in Rooted Trees using Contrastive LearningPreprint
Rasmussen, T. E., & Sørensen, S. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2401.01242
“In this work we propose using the parsimony score to train a binary event encoder using a contrastive approach [6].”
MethodDetecting industrial vehicles’ duty levels using contrastive learning
Cagliero, L., Buccafusco, S., Vaccarino, F., Salvatori, L., & Loti, R. (2023) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata59044.2023.10386964
“…[11]. c) Contrastive Learning: Contrastive Learning (CL) is a particular subclass of Representation Learning, where two samples representing the same pattern should lie close together in the latent space whereas samples corresponding to different patterns should be pushed away from each other [12].”
BackgroundCho, K., Kim, K., Jeong, J., Nam, Y., Kim, J., Choi, C., Lee, S., Hong, G.-S., Seo, J. B., & Kim, N. (2024) · Journal of Imaging Informatics in Medicine · doi:10.1007/s10278-024-01032-x · Cited by 1
Comment on egusphere-2023-2770
Zhang, H., Li, X., Ramelli, F., et al. (2024) · doi:10.5194/egusphere-2023-2770-rc1
How Rapid-Eye-Movement Dreams May Facilitate Learning and CreativityPreprint
Deperrois, N., Petrovici, M. A., Jordan, J., Huber, L. S., & Senn, W. (2024) · Preprints.org · doi:10.20944/preprints202403.0684.v1
Maximal Coding Rate Reduction for Graph Embeddings
Chi, Z., & Gao, J. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10447502 · Cited by 1
“This research is also motivated by the empirical successes of the integration of image learning techniques, including contrastive representation learning[1], into GNN models.”
BackgroundMethodComment on egusphere-2023-2770
Zhang, H., Li, X., Ramelli, F., et al. (2024) · doi:10.5194/egusphere-2023-2770-rc2
Learning explainable constrained representation for time
Amouri, H. E. (2023) · theses.fr (ABES) · doi:10.70675/15168708z0ff6z45bfzadd0z8925f7300d13
Image Restoration for Beautification
Xu, D., Jiang, Y., & Wang, Z. (2023) · Springer eBooks · doi:10.1007/978-3-031-50269-9_15
Hu, Y., Zuo, J., Zhao, A., Iannucci, B., & Joe‐Wong, C. (2024) · doi:10.1109/fmsys62467.2024.00005
“One notable strategy is contrastive representation learning, which improves the model by differentiating among varied data samples [22].”
None-Exemplar Class Incremental Learning with Feature Contrast and Expanded Distillation
Wang, Y., & Zhao, H. (2023) · International Conference on Automation, Robotics and Computer Engineering (ICARCE) · doi:10.1109/icarce59252.2024.10492612
“In the contrastive loss formulation, the target can be dynamically defined in terms of the metric embedding in order to measure the distance, or similarity, between embeddings in the latent space [31].”
BackgroundMethodEnhancing the performance of selectivity estimation with machine learning techniques
Meng, Z. (2024) · DR-NTU (Nanyang Technological University) · doi:10.32657/10356/174777
Zhang, H. (2024) · doi:10.5194/egusphere-2023-2770-ac2
Zhang, H. (2024) · doi:10.5194/egusphere-2023-2770-ac1
Kumar, S. S. (2024) · Journal of Electrical Systems · doi:10.52783/jes.2847
“Using the broad term distributed word representation [23], the feature vectors from the tweets or reviews are mapped.”
BackgroundMulti-Source Fusion User Authentication Based on Channel and Image Features
Wang, A., Tang, J., Luo, H., Wen, H., Han-Ho, P., & Chang, S. Y. (2023) · International Conference on Sensing, Measurement, Communication and Internet of Things Technologies (SMC-IoT) · doi:10.1109/smc-iot62253.2023.00013
Bridging Simulation-based Inference and Hierarchical Modeling : Applications in Neuroscience
Rouillard, L. (2024) · theses.fr (ABES) · doi:10.70675/a47096c5zef68z40f5z92c6zc056f01f10e3
Eskander, M., & Othman, H. H. A. (2024) · IEEE International Conference on Computing and Machine Intelligence (ICMI) · doi:10.1109/icmi60790.2024.10585806
“Data in the input space are then fed into the model, which is divided into two main components: the base encoder and the projection head [4].”
BackgroundMethodInfluentialLinear Transformations in Masked Face Recognition: Bridging the Representation Gap
Ali, E., Elsheikh, A., & Hassanein, M. A. (2023) · International Conference on Computational Science and Computational Intelligence (CSCI) · doi:10.1109/csci62032.2023.00187
“This is part of contrastive learning [17], effective for scenarios with varying conditions.”
BackgroundAudio Classification with Semi-Supervised Contrastive Loss and Consistency Regularization
Xu, J.-W., & Yeh, Y.-R. (2024) · IEEE Annual International Computer Software and Applications Conference (COMPSAC) · doi:10.1109/compsac61105.2024.00279
“For a more comprehensive review, refer to [23].”
BackgroundTwo-Stage Medical Image-Text Transfer with Supervised Contrastive Learning
Wang, X., Yin, S., Wang, Y., Li, J., & Li, S. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72353-7_32
High-Dimensional Feature Fusion for Font Generation with Denoising Diffusion
Liu, X., & Zhang, X. (2024) · International Conference on Machine Learning and Computer Application (ICMLCA) · doi:10.1109/icmlca63499.2024.10754580
Xiong, F., Zhou, Y., Hu, X., & Duan, S. (2024) · Applied Intelligence · doi:10.1007/s10489-024-05957-2
“Its extreme feature extraction capability and taskability are well suited for application on edge devices, and it will be more flexible in handling tasks due to its self-supervised nature [31, 33, 34].”
Vegetarianism Discourse in Russian Social Media: A Case Study
Gorduna, N., & Vanetik, N. (2024) · Applied Sciences · doi:10.3390/app15010259
Huang, Z., & He, K. (2024) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm62325.2024.10822674
“In unsupervised or self-supervised learning, positive samples generally come from data augmentations of the anchor [16].”
Music Emotion Recognition Using Hierarchical Contrastive Learning
Jung, S., Choi, Y., Smith, E. C., & Wang, M. Y. (2024) · Artificial Intelligence x Humanities, Education, and Art (AIxHEART) · doi:10.1109/aixheart62327.2024.00007
Tisha, S. N., & Nallani, G. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-77389-1_20
Contrastive Representation Learning for Regression-Based Aerodynamics Problem: Fuselage and Airfoil
Kamal, T., & Saha, S. (2025) · AIAA SciTech Forum · doi:10.2514/6.2025-0477 · Cited by 1
Feng, X., Li, C., & Zhang, S. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5087850 · Cited by 1
On Contrastive Learning for Domain Generalization
Nguyen, T. D., & Brown, D. R. (2024) · International Conference on Machine Learning and Applications · doi:10.1109/icmla61862.2024.00045
“CL aims to learn a representation function such that similar samples are pulled closer to each other while dissimilar samples are pushed far apart in the representation/feature space [11]–[13].”
Contrastive and Few-shot Learning for Identifying Entanglement in Endangered Marine Species
Asante, B. K. A., & Imamura, H. (2024) · International Conference on Networks, Communication and Computing (ICNCC) · doi:10.1145/3711650.3711656
“Contrastive learning has emerged as a very useful approach in the field of self-supervised learning, especially for learning representations from unlabeled data [9] .”
Research Progress and Prospects of Pre-Training Technology for Electromagnetic Signal Analysis
Lu, X., Li, Q., Long, H., Wang, L., Liang, X., Chen, W., Zhang, X. “., Zeng, X., & Liu, Y. (2025) · IEEE Access · doi:10.1109/access.2025.3553739
“Contrastive learning typically uses [12] or [13] to optimize the model and generate a representation space, with a contrastive learning strategy of pulling each anchor point together with its ‘‘positive’’ samples and separating the anchor points from its ‘‘negative’’ samples.”
Contrastive Learning based on Diffusion Graph for Implicit Collaborative Filtering
Shi, Z., Chen, E., Liu, S., & Wang, B. (2024) · IEEE Smart World Congress (SWC) · doi:10.1109/swc62898.2024.00182
“Contrastive learning (CL) [4] has been widely adopted in solving the OCCF problem [5; 6; 7; 8], such as graph neural network (GNNs)-based models (e.g., PinSage [9] and Light-GCN [10]) and self-supervised learning (SSL)-based models (e.g., SGL [11] and SimGCL [12]).”
Deep Learning and Machine Learning: Contrastive Learning, from scratch to applicationPreprint
Song, X., Chen, K., Bi, Z., Niu, Q., Liu, J., Peng, B., Zhang, S., Liu, M., Li, M., Pan, X., et al. (2025) · OSF Preprints (OSF Preprints) · doi:10.31219/osf.io/7f65r_v3 · Cited by 2
ugli, A. A. A., & ugli, R. S. A. (2025) · INTERNATIONAL JOURNAL OF SCIENCE AND TECHNOLOGY · doi:10.70728/tech.v2.i07.0012
“This highlights a poignant need to explore how multimodal processing frameworks can be adapted to meet the needs of varied communicative contexts, including educational environments and interpersonal interactions [10].”
InfluentialHarper, K., Nasir, W., & Everett, J. (2025) · Preprints.org · doi:10.20944/preprints202505.1275.v1
Fan, Z., Zhang, B., He, M. X., Teng, Z., & Fan, J. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5269808
Veerarajan, R., & Vanitha, V. (2025) · International Journal of Basic and Applied Sciences · doi:10.14419/x1f6be61 · Cited by 1
Unsupervised Text Semantic Matching Based on Contrastive Learning
Chen, C., Zhang, R., Wei, C. Y., & Jing, X. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-6459-7_21
Unveiling the Hidden Pandemic of IoT Malware with Biological and Health Approaches
Craufurd-Hill, J., Kaehler, B. D., Kasmarik, K., & Lynar, T. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-96-1758-6_1
HyPeFL: Tackling Data Heterogeneity via Hypernetwork in Personalized Federated Learning
Chen, Q., Zhao, Y., Li, J., Liu, Y., & Liu, Y. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-7033-8_15
García-Salmerón, J., García, J. M., González‐Férez, P., & Bernabé, G. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5364237
Xia, Z., Yu, X., & Chen, C. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5369762
CREST-Former: A Region-Enhanced Swin Transformer for Pest Recognition Based on Contrastive LearningPreprint
Zou, J., Yang, W., Yin, Y., Feng, Z., & Li, C. (2025) · Research Square · doi:10.21203/rs.3.rs-6644439/v1
A Contrastive Learning and Region-Guided Approach for Long Clinical Text Classification
Lv, S., Li, J., Ouyang, Y., Qi, H., Jiang, Y., & Zhao, Q. Z. (2025) · IEEE Annual International Computer Software and Applications Conference (COMPSAC) · doi:10.1109/compsac65507.2025.00293
“Typically, CL models [21] compute the final projections of positive pairs ( b i 1 , b i 2 ) as z i 1 = F ( h i 1 ) and z i 2 = F ( h i 2 ) , where F is a shared projection model parameterized by θ F and implemented as a 2-layer MLP.”
Subsequence heterogeneity contrastive learning for time series anomaly detection
Zhang, X., Zhang, X., Li, Q., Huang, C., Yin, T., & Yan, M. (2025) · Information Sciences · doi:10.1016/j.ins.2025.122680
Li, J., & Wang, Z. (2025) · International Conference on Artificial Intelligence and Electromechanical Automation (AIEA) · doi:10.1109/aiea66061.2025.11160564
“Contrastive learning [12] represents a self-supervised learning paradigm whose fundamental principle involves training models to recognize feature representations of similar samples through comparative mechanisms.”
Boosting Commit classification with Contrastive LearningPreprint
Tong, J., Wang, Z., & Rui, X. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5547721
Zhou, Y., & Li, W. (2025) · International Conference on Computer Information Science and Application Technology (CISAT) · doi:10.1109/cisat66811.2025.11181896
“Contrastive learning frameworks [15] have been particularly effective in learning invariant features by maximizing agreement between differently augmented views of the same data.”
Building Generalizable Deep Learning Solutions for AIoT Applications
Xu, H. (2025) · International Conference on Mobile Systems, Applications, and Services · doi:10.1145/3711875.3736673
“We further explore sensing scenarios where target data is inaccessible and propose enhancing contrastive learning [10] by injecting additional physics-related information, such as timestamp, user ID, or device ID [2].”
Learning Representations Through Contrastive Neural Model CheckingPreprint
Krsmanovic, V., Cosler, M., Ghanem, M., & Finkbeiner, B. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.01853
“Contrastive learning enables the development of more robust (Xue et al., 2022) and discriminative representations (Le-Khac et al., 2020).”
Metadata
- Authors
- Le-Khac, P. H., Healy, G., & Smeaton, A. F.
- Affiliations
- Dublin City University (IE)
- Venue
- IEEE Access
- Published
- 2020
- Volume
- 8
- Pages
- 193907-193934
- ISSN
- 2169-3536 (ISSN-L)
- OpenAlex
- W3096655658
- Cited by
- 862
- References
- 144
- Citations by year
- 2026: 111 · 2025: 197 · 2024: 212 · 2023: 157 · 2022: 113 · 2021: 71 · 2020: 1
- FWCI
- 33.61
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- IEEE Access (publishedVersion, cc-by)
- arXiv (Cornell University) (acceptedVersion)
- Dublin City University Open Access Institutional Repository (Dublin City University) (submittedVersion, cc-by)
- Arrow@dit (Dublin Institute of Technology) (submittedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- arXiv (Cornell University) (cc-by)
- Funders
- Science Foundation Ireland through the Science Foundation Ireland (SFI) Centre for Research Training in Machine Learning (18/CRT/6183)
- Insight Centre for Data Analytics (SFI/12/RC/2289_P2)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Algorithms
- Keywords
- contrastive learning, representation learning, self-supervised learning, metric learning, inductive bias, computer vision, natural language processing, audio signal processing, reinforcement learning
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.