The Variational Coupled Gaussian Process Dynamical Model

Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10613)


We present a full variational treatment of the Coupled Gaussian Process Dynamical Model (CGPDM) with non-marginalized coupling mappings. The CGPDM generates high-dimensional trajectories from coupled low-dimensional latent dynamical models. The deterministic variational treatment obviates the need for sampling and facilitates the use of the CGPDM on larger data sets. The non-marginalized coupling mappings allow for a flexible exchange of the constituent dynamics models at run time. This exchange possibility is crucial for the construction of modular movement primitive models. We test the model against the marginalized CGPDM, dynamic movement primitives and temporal movement primitives, finding that the CGPDM generally outperforms the other models. Human observers can hardly distinguish CGPDM-generated movements from real human movements.


Gaussian process Variational methods Movement primitives Modularity 



DFG-IRTG 1901 ‘The Brain in Action’, DFG-SFB-TRR 135 project C06. We thank Olaf Haag for help with rendering the movies, and Björn Büdenbender for assistance with MoCap.


  1. 1.
    Bastien, F., Lamblin, P., Pascanu, R., Bergstra, J., Goodfellow, I.J., Bergeron, A., Bouchard, N., Bengio, Y.: Theano: new features and speed improvements. In: Deep Learning and Unsupervised Feature Learning NIPS Workshop (2012)Google Scholar
  2. 2.
    Bauer, M., van der Wilk, M., Rasmussen, C.: Understanding probabilistic sparse Gaussian process approximations. Technical report, arXiv:1606.04820 (2016)
  3. 3.
    Bishop, C.M.: Pattern Recognition and Machine Learning (Information Science and Statistics). Springer, New York (2006)zbMATHGoogle Scholar
  4. 4.
    Bizzi, E., Cheung, V., d’Avella, A., Saltiel, P., Tresch, M.: Combining modules for movement. Brain Res. Rev. 57(1), 125–133 (2008)CrossRefGoogle Scholar
  5. 5.
    Clever, D., Harant, M., Koch, K.H., Mombaur, K., Endres, D.M.: A novel approach for the generation of complex humanoid walking sequences based on a combination of optimal control and learning of movement primitives. Rob. Aut. Sys. 83, 287–298 (2016). doi: 10.1016/j.robot.2016.06.001 CrossRefGoogle Scholar
  6. 6.
    Frigola, R., Chen, Y., Rasmussen, C.: Variational Gaussian process state-space models. In: Ghahramani, Z., Welling, M., Cortes, C., Lawrence, N., Weinberger, K. (eds.) Advances in NIPS, vol. 27, pp. 3680–3688 (2014)Google Scholar
  7. 7.
    Giese, M.A., Mukovskiy, A., Park, A.-N., Omlor, L., Slotine, J.-J.E.: Real-time synthesis of body movements based on learned primitives. In: Cremers, D., Rosenhahn, B., Yuille, A.L., Schmidt, F.R. (eds.) Statistical and Geometrical Approaches to Visual Motion Analysis. LNCS, vol. 5604, pp. 107–127. Springer, Heidelberg (2009). doi: 10.1007/978-3-642-03061-1_6 CrossRefGoogle Scholar
  8. 8.
    Hinton, G.E.: Products of experts. In: Proceedings of ICANN 1999, vol. 1, pp. 1–6 (1999)Google Scholar
  9. 9.
    Ijspeert, A.J., Nakanishi, J., Hoffmann, H., Pastor, P., Schaal, S.: Dynamical movement primitives: learning attractor models for motor behaviors. Neu. Comp. 25(2), 328–373 (2013)CrossRefzbMATHMathSciNetGoogle Scholar
  10. 10.
    Jones, E., Oliphant, T., Peterson, P., et al.: SciPy: Open source scientific tools for Python (2001). Accessed 9 Oct 2015
  11. 11.
    Mattos, C.L.C., Dai, Z., Damianou, A., Forth, J., Barreto, G.A., Lawrence, N.D.: Recurrent Gaussian processes. Technical report, arXiv:1511.06644 (2016)
  12. 12.
    Paraschos, A., Daniel, C., Peters, J., Neumann, G.: Probabilistic movement primitives. In: Burges, C., Bottou, L., Welling, M., Ghahramani, Z., Weinberger, K. (eds.) Advances in NIPS, vol. 26, pp. 2616–2624 (2013)Google Scholar
  13. 13.
    Pastor, P., Kalakrishnan, M., Righetti, L., Schaal, S.: Towards associative skill memories. In: IEEE-RAS International Conference on Humanoid Robots, pp. 309–315 (2012)Google Scholar
  14. 14.
    Sakoe, H., Chiba, S.: Dynamic programming algorithm optimization for spoken word recognition. IEEE Trans. Acoust. Speech Sig. Proc. 26(1), 43–49 (1978)CrossRefzbMATHGoogle Scholar
  15. 15.
    Taubert, N., Christensen, A., Endres, D., Giese, M.: Online simulation of emotional interactive behaviors with hierarchical Gaussian process dynamical models. In: Proceedings of the ACM SAP, pp. 25–32. ACM (2012)Google Scholar
  16. 16.
    Titsias, M.K., Lawrence, N.D.: Bayesian Gaussian process latent variable model. In: Proceedings of the 13th AISTATS, pp. 844–851 (2010)Google Scholar
  17. 17.
    Velychko, D., Endres, D., Taubert, N., Giese, M.A.: Coupling Gaussian process dynamical models with product-of-experts kernels. In: Wermter, S., Weber, C., Duch, W., Honkela, T., Koprinkova-Hristova, P., Magg, S., Palm, G., Villa, A.E.P. (eds.) ICANN 2014. LNCS, vol. 8681, pp. 603–610. Springer, Cham (2014). doi: 10.1007/978-3-319-11179-7_76 Google Scholar
  18. 18.
    Wang, J.M., Fleet, D.J., Hertzmann, A.: Gaussian process dynamical models for human motion. IEEE Trans. Pattern Anal. Mach. Intell. 30(2), 283–298 (2008)CrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  1. 1.Department of PsychologyUniversity of MarburgMarburgGermany

Personalised recommendations