A Transformer-based Framework for Multivariate Time Series Representation Learning
Open accessAbstract
We present a novel framework for multivariate time series representation learning based on the transformer encoder architecture. The framework includes an unsupervised pre-training scheme, which can offer substantial performance benefits over fully supervised learning on downstream tasks, both with but even without leveraging additional unlabeled data, i.e., by reusing the existing data samples. Evaluating our framework on several public multivariate time series datasets from various domains and with diverse characteristics, we demonstrate that it performs significantly better than the best currently available methods for regression and classification, even for datasets which consist of only a few hundred training samples. Given the pronounced interest in unsupervised learning for nearly all domains in the sciences and in industry, these findings represent an important landmark, presenting the first unsupervised method shown to push the limits of state-of-the-art performance for multivariate time series regression and classification.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
The citation lists haven’t been fetched yet
References 32 · Cited by 1,278
Research works printed but not linked: 25
- Iz Beltagy, Matthew E. Peters, and Arman Cohan. Longformer: The Long-Document Transformer. arXiv:2004.05150 [cs], April 2020. URL http://arxiv.org/abs/2004.05150. arXiv: 2004.05150. · arXiv
- Filippo Maria Bianchi, Lorenzo Livi, Karl Øyvind Mikalsen, Michael Kampffmeyer, and Robert Jenssen. Learning representations of multivariate time series with missing data. Pattern Recognition, 96:106973, December 2019. ISSN 0031-3203. doi: 10.1016/j.patcog.2019.106973. · DOI
- T. Brown, B. Mann, Nick Ryder, Melanie Subbiah, J. Kaplan, P. Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, G. Krüger, Tom Henighan, R. Child, Aditya Ramesh, D. Ziegler, Jeffrey Wu, Clemens Winter, Christopher Hesse, Mark Chen, E. Sigler, Mateusz Litwin, Scott Gray, Benjamin Chess, J. Clark, Christopher Berner, Sam McCandlish, A. Radford, Ilya Sutskever, and Dario Amodei. Language models are few-shot learners. ArXiv, abs/2005.14165, 2020. · arXiv
- Zihang Dai, Zhilin Yang, Yiming Yang, Jaime Carbonell, Quoc V. Le, and Ruslan Salakhutdinov. Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context. arXiv:1901.02860 [cs, stat], June 2019. URL http://arxiv.org/abs/1901.02860. arXiv: 1901.02860. · arXiv
- Edward De Brouwer, Jaak Simm, Adam Arany, and Yves Moreau. GRU-ODE-Bayes: Continuous Modeling of Sporadically-Observed Time Series. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d\textquotesingle Alché-Buc, E. Fox, and R. Garnett (eds.), Advances in Neural Information Processing Systems 32, pp. 7379-7390.
- Angus Dempster, Franccois Petitjean, and Geoffrey I. Webb. ROCKET: exceptionally fast and accu- rate time series classification using random convolutional kernels. Data Mining and Knowledge Discovery, 2020. doi: 10.1007/s10618-020-00701-z. · DOI
- Vincent Fortuin, M. Hüser, Francesco Locatello, Heiko Strathmann, and G. Rätsch. SOM-VAE: Interpretable Discrete Representation Learning on Time Series. ICLR, 2019.
- Jean-Yves Franceschi, Aymeric Dieuleveut, and Martin Jaggi. Unsupervised Scalable Represen- tation Learning for Multivariate Time Series. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d\textquotesingle Alché-Buc, E. Fox, and R. Garnett (eds.), Advances in Neural Information Processing Systems 32, pp. 4650-4661.
- Sepp Hochreiter. The Vanishing Gradient Problem During Learning Recurrent Neural Nets and Problem Solutions. Int. J. Uncertain. Fuzziness Knowl.-Based Syst., 6(2):107-116, April 1998. ISSN 0218-4885. doi: 10.1142/S0218488598000094. Place: River Edge, NJ, USA Publisher: World Scientific Publishing Co., Inc. · DOI · Paper page
- Cheng-Zhi Anna Huang, Ashish Vaswani, Jakob Uszkoreit, Ian Simon, Curtis Hawthorne, Noam Shazeer, Andrew M Dai, Matthew D Hoffman, Monica Dinculescu, and Douglas Eck. Music transformer: Generating music with long-term structure. In International Conference on Learning Representations, 2018.
- A. Jansen, M. Plakal, Ratheet Pandya, D. Ellis, Shawn Hershey, Jiayang Liu, R. C. Moore, and R. A. Saurous. Unsupervised Learning of Semantic Audio Representations. 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2018. doi: 10.1109/ICASSP. 2018.8461684. · DOI
- A. Kopf, Vincent Fortuin, Vignesh Ram Somnath, and M. Claassen. Mixture-of-Experts Variational Autoencoder for clustering and generating from similarity-based representations. ICLR 2019, 2019.
- Qi Lei, Jinfeng Yi, R. Vaculín, Lingfei Wu, and I. Dhillon. Similarity Preserving Representation Learning for Time Series Analysis. ArXiv, 2017.
- Shiyang Li, Xiaoyong Jin, Yao Xuan, Xiyou Zhou, Wenhu Chen, Yu-Xiang Wang, and Xifeng Yan. Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting. In Advances in Neural Information Processing Systems, pp. 5243-5253, 2019.
- Benjamin Lucas, Ahmed Shifaz, Charlotte Pelletier, Lachlan O'Neill, Nayyar Zaidi, Bart Goethals, Francois Petitjean, and Geoffrey I. Webb. Proximity Forest: An effective and scalable distance- based classifier for time series. Data Mining and Knowledge Discovery, 33(3):607-635, May 2019. ISSN 1384-5810, 1573-756X. doi: 10.1007/s10618-019-00617-3. arXiv: 1808.10594. · DOI
- Xinrui Lyu, Matthias Hueser, Stephanie L. Hyland, George Zerveas, and Gunnar Raetsch. Improving Clinical Predictions through Unsupervised Time Series Representation Learning. In Proceedings of the NeurIPS 2018 Workshop on Machine Learning for Health, 2018. eprint: 1812.00490.
- J. Ma, Zheng Shou, Alireza Zareian, Hassan Mansour, A. Vetro, and S. Chang. Cdsa: Cross-dimensional self-attention for multivariate, geo-tagged time series imputation. ArXiv, abs/1905.09904, 2019. · arXiv
- Razvan Pascanu, Tomas Mikolov, and Yoshua Bengio. On the difficulty of training recurrent neural networks. In Proceedings of the 30th International Conference on International Conference on Machine Learning -Volume 28, ICML'13, pp. III-1310-III-1318, Atlanta, GA, USA, June 2013. JMLR.org.
- Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, W. Li, and Peter J. Liu. Exploring the Limits of Transfer Learning with a Unified Text-to- Text Transformer. ArXiv, abs/1910.10683, 2019. · arXiv
- Sheng Shen, Zhewei Yao, Amir Gholami, Michael W. Mahoney, and Kurt Keutzer. PowerNorm: Rethinking Batch Normalization in Transformers. arXiv:2003.07845 [cs], June 2020. arXiv: 2003.07845. · arXiv
- Ahmed Shifaz, Charlotte Pelletier, F. Petitjean, and Geoffrey I. Webb. TS-CHIEF: a scalable and accurate forest algorithm for time series classification. Data Mining and Knowledge Discovery, 2020. doi: 10.1007/s10618-020-00679-8. · DOI
- C. Tan, C. Bergmeir, Franc ¸ois Petitjean, and Geoffrey I. Webb. Monash University, UEA, UCR Time Series Regression Archive. ArXiv, 2020a.
- Chang Wei Tan, Christoph Bergmeir, Francois Petitjean, and Geoffrey I Webb. Time series regres- sion. arXiv preprint arXiv:2006.12672, 2020b. · arXiv
- Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is All you Need. In I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett (eds.), Advances in Neural Information Processing Systems 30, pp. 5998-6008.
- Chuxu Zhang, Dongjin Song, Yuncong Chen, Xinyang Feng, C. Lumezanu, Wei Cheng, Jingchao Ni, B. Zong, H. Chen, and Nitesh V. Chawla. A Deep Neural Network for Unsupervised Anomaly Detection and Diagnosis in Multivariate Time Series Data. In AAAI, 2019. doi: 10.1609/aaai. v33i01.33011409. · DOI
- Authors
- Zerveas, G., Jayaraman, S., Patel, D., Bhamidipaty, A., & Eickhoff, C.
- Affiliations
- Brown University (US)
- IBM (United States) (US)
- Venue
- Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining
- Publisher
- ACM
- Published
- 2021-08-14
- Pages
- 2114-2124
- OpenAlex
- W3092172514
- Cited by
- 1,278
- References
- 32
- Citations by year
- 2026: 254 · 2025: 372 · 2024: 287 · 2023: 261 · 2022: 91 · 2021: 13
- FWCI
- 99.23
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- NSF (National Science Foundation) (IIS-1956221)
- Onassis Foundation
- Topics
- Physical Sciences › Computer Science › Signal Processing › Time Series Analysis and Forecasting
- Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Keywords
- multivariate time series, representation learning, Transformer encoder, unsupervised pre-training, time series regression, time series classification, small training samples, unlabeled data
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
