Time-Series Representation Learning via Temporal and Contextual Contrasting
Open accessThe DOI matches its Crossref record; 28 of the 28 references it cites were found in a registry or an index; 9 of the 21 sources its claims cite were checked for support.
Abstract
Learning decent representations from unlabeled time-series data with temporal dynamics is a very challenging task. In this paper, we propose an unsupervised Time-Series representation learning framework via Temporal and Contextual Contrasting (TS-TCC), to learn time-series representation from unlabeled data. First, the raw time-series data are transformed into two different yet correlated views by using weak and strong augmentations. Second, we propose a novel temporal contrasting module to learn robust temporal representations by designing a tough cross-view prediction task. Last, to further learn discriminative representations, we propose a contextual contrasting module built upon the contexts from the temporal contrasting module. It attempts to maximize the similarity among different contexts of the same sample while minimizing similarity among contexts of different samples. Experiments have been carried out on three real-world time-series datasets. The results manifest that training a linear classifier on top of the features learned by our proposed TS-TCC performs comparably with the supervised training. Additionally, our proposed TS-TCC shows high efficiency in few-labeled data and transfer learning scenarios. The code is publicly available at https://github.com/emadeldeen24/TS-TCC.
Overview
CoChat Summary
TL;DR
TS-TCC is an unsupervised time-series representation learning framework that combines weak/strong augmentations, a cross-view temporal contrasting module, and a contextual contrasting module. It learns representations that perform comparably to supervised training with a linear classifier and are effective in few-label and transfer settings.
Key findings
- On linear evaluation, TS-TCC achieved HAR ACC 90.37 ± 0.34 / MF1 90.38 ± 0.39, Sleep-EDF ACC 83.00 ± 0.71 / MF1 73.57 ± 0.74, and Epilepsy ACC 97.23 ± 0.10 / MF1 95.54 ± 0.08, outperforming SSL-ECG, CPC, and SimCLR and matching or exceeding supervised training on two of three datasets.
- In semi-supervised settings, TS-TCC fine-tuning reached around 70% and 90% for HAR and Epilepsy with only 1% labeled data, and with 10% labeled data achieved performance comparable to supervised training with 100% labeled data on the three datasets.
- In fault-diagnosis transfer learning, TS-TCC fine-tuning averaged 67.82% accuracy versus 63.83% for supervised pretraining, outperforming supervised in 8 of 12 cross-domain scenarios and improving by about 4% overall.
- Ablation showed the cross-view prediction task improved performance by more than 5% on HAR and about 1% on Sleep-EDF and Epilepsy; adding contextual contrasting further improved results.
Methods
- The framework creates two views per sample using weak augmentation (jitter-and-scale) and strong augmentation (permutation-and-jitter), encodes them with a 3-block convolutional encoder, and applies a Transformer-based autoregressive temporal contrasting module with a cross-view prediction task plus a contextual contrasting loss.
- Experiments used UCI HAR, Sleep-EDF, Epilepsy, and a fault diagnosis dataset; data were split 60%/20%/20% for train/validation/test, with subject-wise split for Sleep-EDF, and experiments were repeated 5 times with 5 seeds.
- Evaluation included linear classifier on frozen features, semi-supervised training with 1%, 5%, 10%, 50%, and 75% labeled data, cross-domain transfer learning, ablation, and sensitivity analysis; training used 40 epochs, Adam with lr 3e-4, batch size 128 (32 for few-label), and dataset-specific augmentation/Transformer hyperparameters.
Limitations
- Not stated in the paper: the paper does not include an explicit limitations section.
- Evident from the text: augmentation hyperparameters must be chosen carefully according to the nature of the time-series data, and the main evaluation covers three time-series datasets plus one fault-diagnosis transfer dataset.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures






Verification
Claims the paper makes
8 Valid · 2 Not valid · 5 Not checked · 43 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“However, the pretext tasks can limit the generality of the learned representations.”
What the sources say
- Not checked — no full text of the sourcecites Noroozi and Favaro, 2016 source
- Partly supportedcites Gidaris et al., 2018 sourceExperimental — accuracy not yet measured
The Introduction says representations learned through earlier self-supervised pretext tasks had not matched supervised representations in performance, which supports a qualified concern about their effectiveness. It also calls them good alternatives for transfer to other vision tasks; the paper does not establish that the pretext tasks themselves limit the representations’ generality. No page number is available in our copy. (LLM, conf=0.72)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“First, they may not be able to address the temporal dependencies of data, which are key characteristics of time-series [Franceschi et al. , 2019].”
What the sources say
- Partly supportedcites Franceschi et al., 2019 sourceExperimental — accuracy not yet measured
Section 3, p. 3, uses time-based sampling and an unsupervised time-based criterion for learning time-series representations, supporting the importance of temporal information. But neither that section nor the discussion of other domains in Related Work (p. 2) says that image-based contrastive methods may fail to address temporal dependencies. (LLM, conf=0.76)
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“However, these image-based contrastive learning methods are not able to work well on time-series data for the following reasons.”
What the sources say
- Could not verifycites Franceschi et al., 2019 sourceExperimental — accuracy not yet measured
Related Work (p. 2) says triplet losses have seen little use for time series and identifies limitations of certain existing time-series methods. Unsupervised Training (p. 3) explains why the authors adapt a triplet loss to time series, but neither passage assesses whether image-based contrastive learning methods work well on time-series data. Searches for image-based contrastive methods and their suitability for time series found no passage addressing that claim. (LLM, conf=0.70)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“However, the above two methods are proposed for specific applications and they are not generalizable to other time-series data.”
What the sources say
- Not checked — no full text of the sourcecites Mohsenvand et al., 2020
- Partly supportedcites Cheng et al., 2020 sourceExperimental — accuracy not yet measured
Our copy's Related Work (Sec. 2, p. 2) does characterize prior self-supervised/contrastive work as tied to particular modalities or applications — pretext tasks shown effective for EEG [8] and ECG [9], and Hyvärinen & Morioka's time-contrastive learning [14] 'demonstrated for magnetoencephalography (MEG)' — and the Introduction (Sec. 1, pp. 1-2) adds that applying contrastive learning to biosignals 'requires careful design of appropriate data transformations and consideration of intersubject variability' and that it is 'unclear whether the same techniques used to learn representations for language and vision domains are effective for biosignals'. That supports only the 'proposed for specific applications' half. The categorical part of the claim — that existing EEG/ECG contrastive methods 'are not generalizable to other time-series data' — is nowhere asserted; the paper instead hedges about transfer in the opposite direction (vision/language → biosignals) and states in its Conclusion (Sec. 5, p. 8) that its own subject-aware contrastive approach 'can be applied to other biosignals, such as signals from the eye (EOG) or muscles (EMG)'. So the citing sentence overstates the source by turning a hedged, domain-specificity motivation into a blanket non-generalizability verdict. (LLM, conf=0.68)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Additionally, our proposed TS-TCC shows high efficiency in few-labeled data and transfer learning scenarios.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
32 more claims
Create a free account to see all 58 claims and every source verdict.
- Probably not a substantial claim. 12
- It may state two separate claims as one. 12
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 says what the claim attributes to it.
In the citing paper
“In the time domain, TS-TCC [116] and its extended version CA-TCC [117] designed two time series data augmentation techniques, one is strong augmentation (permutation-and-jitter), and the other is weak augmentation (jitter-and-scale).”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [116]
The cited source says what the claim attributes to it.
In the citing paper
“TS-TCC [116] and its extended version CA-TCC [117] designed a cross prediction task, which uses the context of x T 1 to predict the target in x T 2 , and vice versa uses the context of x T 2 to predict the target in x T 1 .”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [116]
The cited source says what the claim attributes to it.
In the citing paper
“However, CoST and BTSF do not modify the frequency representation, while TF-C [125] directly augments the time series data through frequency perturbations, which has achieved better performance than TS2Vec [118] and TS-TCC [116].”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [116]
The cited source says what the claim attributes to it.
In the citing paper
“Additionally, most SSCL-based classification models emphasize that data augmentation also has a direct impact on the final classification accuracy [116], [118], [122].”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [116]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 3.2 Temporal Contrasting · p. 3
“The contrastive loss tries to minimize the dot product between the predicted representation and the true one of the same sample, while maximizing the dot product with the other samples N t,k within the minibatch.”
§ 3.1 Time-Series Data Augmentation · p. 3
“Contrastive methods try to maximize the similarity among different views of the same sample, while minimizing its similarity with other samples.”
Citations
Cites
Vaswani, 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
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
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
Momentum Contrast for Unsupervised Visual Representation Learning
He, K., Fan, H., Wu, Y., Xie, S., & Girshick, R. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00975 · Cited by 13,347
A simple framework for contrastive learning of visual representations
Chen, T., Kornblith, S., Norouzi, M., & Hinton, G. (2020) · arXiv · doi:10.48550/arxiv.2002.05709 · Cited by 10,857
Representation Learning with Contrastive Predictive Coding
Oord, A. V. D., Li, Y., & Vinyals, O. (2018) · arXiv · doi:10.48550/arxiv.1807.03748 · Cited by 6,242
Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning
Grill, J.-B., Strub, F., Altché, F., Tallec, C., Richemond, P. H., Buchatskaya, E., Doersch, C., Pires, B. A., Guo, Z. D., Azar, M. G., et al. (2020) · arXiv · doi:10.48550/arxiv.2006.07733 · Cited by 3,846
Exploring Simple Siamese Representation Learning
Chen, X., & He, K. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.01549 · Cited by 3,725
Zhang, R., Isola, P., & Efros, A. A. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46487-9_40 · Cited by 3,178
Research works printed but not linked: 1
- [Aggarwal et al. , 2019] Karan Aggarwal, Shafiq R. Joty, Luis Fern´ andez-Luque, and Jaideep Srivastava. Adversarial unsupervised representation learning for activity timeseries. AAAI , 2019.
Cited by
Diffusion Models: A Comprehensive Survey of Methods and Applications
Yang, L., Zhang, Z., Song, Y., Hong, S., Xu, R., Zhao, Y., Zhang, W., Cui, B., & Yang, M.-H. (2024) · ACM Computing Surveys · doi:10.1145/3626235 · Cited by 1,704
“Time series data are widely used with many important real-world applications [54, 152, 228, 243].”
BackgroundTS2Vec: Towards Universal Representation of Time Series
Yue, Z., Wang, Y., Duan, J., Yang, T., Huang, C., Tong, Y., & Xu, B. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i8.20881 · Cited by 850
“TS-TCC (Eldele et al. 2021) encourages consistency of different data augmentations.”
BackgroundMethodInfluentialTSMixer: Lightweight MLP-Mixer Model for Multivariate Time Series Forecasting
Ekambaram, V., Jati, A., Nguyen, N., Sinthong, P., & Kalagnanam, J. (2023) · Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining · doi:10.1145/3580305.3599533 · Cited by 278
Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects
Zhang, K., Wen, Q., Zhang, C., Cai, R., Jin, M., Liu, Y., Zhang, J. Y., Liang, Y., Pang, G., Song, D., et al. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3387317 · Cited by 295
“TS-TCC [116] and its extended version CA-TCC [117] designed a cross prediction task, which uses the context of x T 1 to predict the target in x T 2 , and vice versa uses the context of x T 2 to predict the target in x T 1 .”
MethodInfluentialDeep Learning for Time Series Classification and Extrinsic Regression: A Current Survey
Mohammadi Foumani, N., Miller, L., Tan, C. W., Webb, G. I., Forestier, G., & Salehi, M. (2024) · ACM Computing Surveys · doi:10.1145/3649448 · Cited by 268
“The model’s accuracy was also improved by pre-training the parameters on massive unlabeled data.”
Background
Metadata
- Authors
- Eldele, E., Ragab, M., Chen, Z., Wu, M., Kwoh, C. K., Li, X., & Guan, C.
- Affiliations
- Nanyang Technological University (SG)
- Agency for Science, Technology and Research (SG)
- Venue
- Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence
- Published
- 2021-08
- Pages
- 2352-2359
- OpenAlex
- W3190152617
- Cited by
- 696
- References
- 25
- Citations by year
- 2026: 152 · 2025: 213 · 2024: 181 · 2023: 117 · 2022: 31 · 2021: 2
- FWCI
- 50.23
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- International Joint Conference on Artificial Intelligence (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- DR-NTU (Nanyang Technological University) (submittedVersion)
- arXiv (Cornell University) (cc-by)
- Funders
- Agency for Science, Technology and Research (A20H6b0151, C210112046)
- Topics
- Physical Sciences › Computer Science › Signal Processing › Time Series Analysis and Forecasting
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Physical Sciences › Computer Science › Artificial Intelligence › Advanced Graph Neural Networks
- Keywords
- time series representation learning, unsupervised learning, contrastive learning, data augmentation, temporal dynamics, limited labeled data, transfer 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.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Time-Series Representation Learning via Temporal and Contextual Contrasting. https://app.cochat.ai/papers/10.24963/ijcai.2021/324
- The paper
- Eldele, E., Ragab, M., Chen, Z., Wu, M., Kwoh, C. K., Li, X., & Guan, C. (2021). Time-Series Representation Learning via Temporal and Contextual Contrasting. Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence. https://doi.org/10.24963/ijcai.2021/324
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.