Skip to main content
SpringerLink
Log in

Estimating time series averages from latent space of multi-tasking neural networks

  • Regular Paper
  • Published:
Knowledge and Information Systems Aims and scope Submit manuscript

A Correction to this article was published on 19 September 2023

This article has been updated

Abstract

Time series averages are one key input to temporal data mining techniques such as classification, clustering, forecasting, etc. In practice, the optimality of estimated averages often impacts the performance of such temporal data mining techniques. Practically, an estimated average is presumed to be optimal if it minimizes the discrepancy between itself and members of an averaged set while preserving descriptive shapes. However, estimating an average under such constraints is often not trivial due to temporal shifts. To this end, all pioneering averaging techniques propose to align averaged series before estimating an average. Practically, the alignment gets performed to transform the averaged series, such that, after the transformation, they get registered to their arithmetic mean. However, in practice, most proposed alignment techniques often introduce additional challenges. For instance, Dynamic Time Warping (DTW)-based alignment techniques make the average estimation process non-smooth, non-convex, and computationally demanding. With such observation in mind, we approach time series averaging as a generative problem. Thus, we propose to mimic the effects of temporal alignment in the latent space of multi-tasking neural networks. We also propose to estimate (augment) time domain averages from the latent space representations. With this approach, we provide state-of-the-art latent space registration. Moreover, we provide time domain estimations that are better than the estimates generated by some pioneering averaging techniques.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13
Fig. 14
Fig. 15

Similar content being viewed by others

Change history

Notes

  1. The Python implementation of the proposed architectures, the training setup, and 1NCC evaluation can be found at the following github repository: https://github.com/tsegaterefe/Estimating-Time-Series-Averages-from-Latent-Space-of-Multi-tasking-Neural-Networks/tree/main.

References

  1. Aghabozorgi S, Shirkhorshidi AS, Wah TY (2015) Time-series clustering: a decade review. Inf Syst 53:16–38

    Article  Google Scholar 

  2. Bagnall A, Davis L, Hills J, Lines J (2012) Transformation based ensembles for time series classification. In: Proceedings of the 2012 SIAM international conference on data mining. Society for Industrial and Applied Mathematics, Anaheim,CA, USA, pp 307–318

  3. Bagnall A, Lines J (2014) An experimental evaluation of nearest neighbour time series classification. Technical report, University of East Angelina arXiv:1406.4757

  4. Bock H-H (2008) Origins and extensions of the -means algorithm in cluster analysis. Journal Électronique d’Histoire des Probabilités et de la Statistique [electronic only] 4:1–18

    MathSciNet  Google Scholar 

  5. Bulteau L, Froese V, Niedermeier R (2020) Tight hardness results for consensus problems on circular strings and time series. SIAM J. Discrete Math. 34(3):1854–1883

    Article  MathSciNet  MATH  Google Scholar 

  6. Chen C, Srivastava A (2021) Srvfregnet: elastic function registration using deep neural networks. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (CVPR) workshops. IEEE Computer Society, New Orleans, Louisiana, USA, pp 4462–4471

  7. Chen Y, Keogh E, Hu B, Begum N, Bagnall A, Mueen A, Batista G (2015) The UCR time series classification archive. www.cs.ucr.edu/~eamonn/time_series_data/

  8. Christian S, Liu W, Jia Y, Pierre S, Scott R, Dragomir A, Dumitru E, Vincent V, Andrew R (2015) Going deeper with convolutions. In: 2015 IEEE conference on computer vision and pattern recognition (CVPR). IEEE Computer Society, Boston, MA, USA, pp 1–9

  9. Cuturi M, Blondel M (2017) Soft-dtw: a differentiable loss function for time-series. In: Proceedings of the 34th international conference on machine learning. JMLR.org, Sydney, NSW, Australia, pp 894–903

  10. Debella TT, Shawel BS, Devanne M, Weber J, Woldegebreal DH, Pollin S, Forestier G (2022) Deep representation learning for cluster-level time series forecasting. In: 8th International conference on time series and forecasting (ITISE). MDPI, Gran Canaria, Spain, pp 1–11

  11. Demšar J (2006) Statistical comparisons of classifiers over multiple data sets. J Mach Learn Res 7:1–30

    MathSciNet  MATH  Google Scholar 

  12. Der Maaten LV, Hinton G (2008) Visualizing data using t-SNE. J Mach Learn Res 9:2579–2605

  13. Detlefsen NS, Freifeld O, Hauberg S (2018) Deep diffeomorphic transformer networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR). In: IEEE Computer Society, Lake City, UT, USA, pp 4403–4412

  14. Diederik PK, Max W (2014) Auto-encoding variational bayes. In: 2nd International conference on learning representations (ICLR 2014). ICLR, Banff, AB, Canada, pp 1–14

  15. Dong G, Liao G, Liu H, Kuang G (2018) A review of the autoencoder and its variants: a comparative perspective from target recognition in synthetic-aperture radar images. IEEE Geosci Remote Sens Mag 6(3):44–68

    Article  Google Scholar 

  16. Fawaz HI, Benjamin L, Forestier G, Charlotte P, Schmidt DF, Weber J, Webb GI, Idoumghar L, Muller PA, Petitjean F (2020) Inceptiontime: finding alexnet for time series classification. Data Min Knowl Disc 34:1936–1962

    Article  MathSciNet  Google Scholar 

  17. Fawaz HI, Forestier G, Weber J, Idoumghar L, Alain-Muller P (2019) Deep learning for time series classification: a review. Data Min Knowl Disc 33(4):917–963

    Article  MathSciNet  MATH  Google Scholar 

  18. Gee AH, Garcia-Olano D, Ghosh J, Paydarfar D (2019) Explaining deep classification of time-series data with learned prototypes. CEUR Workshop Proc 2429:15–22

    Google Scholar 

  19. Glorot X, Bengio Y (2010) Understanding the difficulty of training deep feedforward neural networks. In: Proceedings of the 13th conference on artificial intelligence and statistics. PMLR, Chia Laguna Resort, Sardinia, Italy, pp 249–256

  20. Goodfellow I, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, Courville A, Bengio Y (2014) Generative adversarial networks, pp 1–9

  21. Gupta L, Molfese D, Tammana R, Simos P (1996) Nonlinear alignment and averaging for estimating the evoked potential. IEEE Trans Biomed Eng 43(4):348–356

    Article  Google Scholar 

  22. He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In: Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR). IEEE Computer Society, Las Vegas, NV, USA, pp 770–778

  23. Itakura F (1975) Minimum prediction residual principle applied to speech recognition. IEEE Trans Acoust Speech Signal Process 23(1):67–72

    Article  Google Scholar 

  24. Iwana BK, Uchida S (2021) An empirical survey of data augmentation for time series classification with neural networks. PLOS ONE 16(7):1–32

    Article  Google Scholar 

  25. Jain B, Froese V, Schultz D (2019) An average-compress algorithm for the sample mean problem under dynamic time warping. CoRR arXiv:abs/1909.13541

  26. Jain J, Schultz D (2018) Asymmetric learning vector quantization for efficient nearest neighbor classification in dynamic time warping spaces. Pattern Recogn 76:349–366

    Article  Google Scholar 

  27. Junyuan X, Ross G, Ali F (2016) Unsupervised deep embedding for clustering analysis. In: Proceedings of the 33rd international conference on international conference on machine learning, vol 48, pp 478–487

  28. Kaiming H, Xiangyu Z, Shaoqing R, Jian S (2015) Delving deep into rectifiers: surpassing human-level performance on imagenet classification. In: 2015 IEEE international conference on computer vision (ICCV). IEEE Computer Society, Santiago, Chile, pp 1026–1034

  29. Kowsar Y, Moshtaghi M, Velloso E, Bezdek JC, Kulik L, Leckie C (2022) Shape-sphere: a metric space for analysing time series by their shape. Inf Sci 582:198–214

    Article  MathSciNet  Google Scholar 

  30. Krizhevsky A, Sutskever I, Hinton GE (2012) Imagenet classification with deep convolutional neural networks. In: Advances in neural information processing systems. Neural Information Processing Systems Foundation Inc. (NeurIPS), Lake Tahoe, Nevada, USA, pp 1106–1114

  31. Lafabregue B, Weber J, Gançarski P, Forestier G (2021a) End-to-end deep representation learning for time series clustering: a comparative study. In: Data mining and knowledge discovery, pp 1–53

  32. Lafabregue B, Weber J, Gançarski P, Forestier G (2021) End-to-end deep representation learning for time series clustering: a comparative study. Data Min Knowl Disc 36:29–81

    Article  MathSciNet  MATH  Google Scholar 

  33. Lin J, Keogh E, Wei L, Lonardi S (2007) Experiencing sax: a novel symbolic representation of time series. Data Min Knowl Disc 15(2):107–144

    Article  MathSciNet  Google Scholar 

  34. Lin J, Li Y (2009) Finding structural similarity in time series data using bag-of-patterns representation. In: International conference on scientific and statistical database management. Springer, New Orleans, LA, USA, pp 461–477

  35. Lines J (2015) Time series classification through transformation and ensembles. Ph.D. Thesis, School of Electrical and Computer Engineering, University of East Anglia

  36. Niennattrakul V, Ratanamahatana CA (2009) Shape averaging under time warping. In: 2009 6th international conference on electrical engineering/electronics, computer, telecommunications and information technology. IEEE, Chonburi, Thailand, pp 626–629

  37. Niennattrakul V, Srisai D, Ratanamahatana CA (2012) Shape-based template matching for time series data. Knowl Based Syst 26:1–8

    Article  Google Scholar 

  38. Ongwattanakul S, Srisai D (2009) Contrast enhanced dynamic time warping distance for time series shape averaging classification. In: Proceedings of the 2nd international conference on interaction sciences: information technology, culture and human. Association for Computing Machinery, pp 976–981

  39. Paparrizos J, Gravano L (2015) K-shape: Efficient and accurate clustering of time series. In: Proceedings of the 2015 ACM SIGMOD international conference on management of data. Association for Computing Machinery, Melbourne, Victoria, Australia, pp 1855–1870

  40. Petitjean F, Gançarski P (2012) Summarizing a set of time series by averaging: From Steiner sequence to compact multiple alignment. Theoret Comput Sci 414(1):76–91

    Article  MathSciNet  MATH  Google Scholar 

  41. Petitjean F, Ketterlin A, Gançarski P (2011) A global averaging method for dynamic time warping, with applications to clustering. Pattern Recogn 44(3):678–693

    Article  MATH  Google Scholar 

  42. Ruiz EV, Casacuberta Nolla F, Segovia HR (1985) Is the DTW “distance’’ really a metric? An algorithm reducing the number of DTW comparisons in isolated word recognition. Speech Commun 4(4):333–344

    Article  Google Scholar 

  43. Sakoe H, Chiba S (1978) Dynamic programming algorithm optimization for spoken word recognition. IEEE Trans Acoust Speech Signal Process 26(1):43–49

    Article  MATH  Google Scholar 

  44. Salvador S, Chan P (2007) Toward accurate dynamic time warping in linear time and space. Intell Data Anal 11(5):561–580

    Article  Google Scholar 

  45. Schultz D, Jain B (2018) Nonsmooth analysis and subgradient methods for averaging in dynamic time warping spaces. Pattern Recogn 74:340–358

    Article  Google Scholar 

  46. Shapira Weber RA, Eyal M, Skafte N, Shriki O, Freifeld O (2019a) Diffeomorphic temporal alignment nets. In: Advances in neural information processing systems 32 (NeurIPS 2019). Neural Information Processing Systems Foundation Inc. (NeurIPS), Vancouver,Canada, pp 6574–6585. http://papers.nips.cc/paper/8884-diffeomorphic-temporal-alignment-nets.pdf

  47. Shapira Weber RA, Eyal M, Skafte N, Shriki O, Freifeld O (2019b) Diffeomorphic temporal alignment nets: supplementary material. In: Advances in neural information processing systems 32 (NeurIPS 2019). Neural Information Processing Systems Foundation Inc. (NeurIPS), Vancouver,Canada, pp 6574–6585

  48. Shawel BS, Debella TT, Tesfaye G, Tefera YY, Woldegebreal DH (2020) Hybrid prediction model for mobile data traffic: a cluster-level approach. In: 2020 International joint conference on neural networks (IJCNN). IEEE, Glasgow, UK, pp 1–8

  49. Simonyan K, Zisserman A (2015) Very deep convolutional networks for large-scale image recognition. In: 3rd International conference on learning representations (ICLR). ICLR, San Diego, CA, USA, pp 1–14

  50. Srisai D, Ratanamahatana CA (2009) Efficient time series classification under template matching using time warping alignment. In: 2009 Fourth international conference on computer sciences and convergence information technology. IEEE, Seoul, Korea, pp 685–690

  51. Srivastava A, Klassen EP (2016) Functional and shape data analysis, vol 1. Springer, NY

  52. Tavenard R, Faouzi J, Vandewiele G, Divo F, Androz G, Holtz C, Payne M, Yurchak R, Rußwurm M, Kolar K, Woods E (2020) Tslearn, a machine learning toolkit for time series data. J Mach Learn Res 21(118):1–6

    Google Scholar 

  53. Terefe T, Devanne M, Weber J, Hailemariam D, Forestier G (2020) Time series averaging using multi-tasking autoencoder. In 2020 IEEE 32nd international conference on tools with artificial intelligence (ICTAI). IEEE Computer Society, Baltimore, MD, USA, pp 1065–1072

  54. Wei W (2006) Time series analysis: univariate and multivariate methods, 2nd edn. Pearson Addison Wesley, New York

    MATH  Google Scholar 

  55. Xie J, Girshick R, Farhadi A (2016) Unsupervised deep embedding for clustering analysis. In: Proceedings of the 33rd international conference on international conference on machine learning (ICML’16). ICML, NY, USA, pp 478–487

  56. Ye L, Keogh E (2009) Time series shapelets: A new primitive for data mining. In: Proceedings of the 15th ACM SIGKDD international conference on knowledge discovery and data mining. Association for Computing Machinery, Paris, France, pp 947–956

Download references

Acknowledgements

The authors would like to thank the creators and providers of the UCR datasets: Hoang Anh Dau, Anthony Bagnall, Kaveh Kamgar, Chin-Chia Michael Yeh, Yan Zhu, Shaghayegh Gharghabi, Chotirat Ann Ratanamahatana, Eamonn Keogh and Mustafa Baydogan. Moreover, the authors would also like to thank the university of Strasbourg for allowing us to use its HPC clusters (Mesocentrer). Last but not least, we would also like to thank Ouloufa Dorani and Sophia Nicée, and Dr. Esayas Gebreyouhannes for their roles in the continuation of the Ethio-France Ph.D. program under challenging circumstances. This work got conducted under the support of the French embassy for the African Union and Ethiopia and the former Ethiopian Ministry of Science and Higher Education (MOSHE).

Author information

Authors and Affiliations

  1. School of Electrical and Computer Engineering, Addis Ababa Institute of Technology, Addis Ababa University, King George VI Street, P.O. Box 385, Addis Ababa, Ethiopia

    Tsegamlak Terefe & Dereje Hailemariam

  2. Institut de Recherche en Informatique, Mathématiques, Automatique et Signal, Université de Haute-Alsace, Mulhouse, France

    Tsegamlak Terefe, Maxime Devanne, Jonathan Weber & Germain Forestier

Authors
  1. Tsegamlak Terefe
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Maxime Devanne
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Jonathan Weber
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Dereje Hailemariam
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Germain Forestier
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Tsegamlak Terefe.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Terefe, T., Devanne, M., Weber, J. et al. Estimating time series averages from latent space of multi-tasking neural networks. Knowl Inf Syst 65, 4967–5004 (2023). https://doi.org/10.1007/s10115-023-01927-1

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10115-023-01927-1

Keywords

Search

Navigation