Spacetime Autoencoders Using Local Causal States

Author(s): Rupe, Adam; Crutchfield, James P | Abstract: Local causal states are latent representations that capture organized pattern and structure in complex spatiotemporal systems. We expand their functionality, framing them as spacetime autoencoders. Previously, they were only considered as maps from observable spacetime fields to latent local causal state fields. Here, we show that there is a stochastic decoding that maps back from the latent fields to observable fields. Furthermore, their Markovian properties define a stochastic dynamic in the latent space. Combined with stochastic decoding, this gives a new method for forecasting spacetime fields.

[1]  Steven L. Brunton,et al.  Deep learning for universal linear embeddings of nonlinear dynamics , 2017, Nature Communications.

[2]  Frank Noé,et al.  Kernel methods for detecting coherent structures in dynamical data. , 2019, Chaos.

[3]  Jaideep Pathak,et al.  Model-Free Prediction of Large Spatiotemporally Chaotic Systems from Data: A Reservoir Computing Approach. , 2018, Physical review letters.

[4]  B. R. Noack.  Turbulence, Coherent Structures, Dynamical Systems and Symmetry , 2013 .

[5]  I. Mezić,et al.  Analysis of Fluid Flows via Spectral Properties of the Koopman Operator , 2013 .

[6]  Georg M. Goerg,et al.  LICORS: Light Cone Reconstruction of States for Non-parametric Forecasting of Spatio-Temporal Systems , 2012, 1206.2398.

[7]  James Odell,et al.  Between order and chaos , 2011, Nature Physics.

[8]  D. Giannakis,et al.  Operator-theoretic framework for forecasting nonlinear time series with kernel analog techniques , 2019, Physica D: Nonlinear Phenomena.

[9]  Hao Wu,et al.  VAMPnets for deep learning of molecular kinetics , 2017, Nature Communications.

[10]  Kyle Johnson,et al.  Coupled Map Lattices as Musical Instruments , 2018, Computer Music Journal.

[11]  Clarence W. Rowley,et al.  Linearly-Recurrent Autoencoder Networks for Learning Dynamics , 2017, SIAM J. Appl. Dyn. Syst..

[12]  James P. Crutchfield,et al.  Equations of Motion from a Data Series , 1987, Complex Syst..

[13]  Petros Koumoutsakos,et al.  Machine Learning for Fluid Mechanics , 2019, Annual Review of Fluid Mechanics.

[14]  Kurt Hornik,et al.  Neural networks and principal component analysis: Learning from examples without local minima , 1989, Neural Networks.

[15]  Vipin Kumar,et al.  Integrating Physics-Based Modeling with Machine Learning: A Survey , 2020, ArXiv.

[16]  Adam Rupe,et al.  DisCo: Physics-Based Unsupervised Discovery of Coherent Structures in Spatiotemporal Systems , 2019, 2019 IEEE/ACM Workshop on Machine Learning in High Performance Computing Environments (MLHPC).

[17]  J. Crutchfield.  Between order and chaos , 2011, Nature Physics.

[18]  Mohammad M. Sultan,et al.  Variational encoding of complex dynamics. , 2017, Physical review. E.

[19]  G. G. Stokes.  "J." , 1890, The New Yale Book of Quotations.

[20]  James P. Crutchfield,et al.  Phenomenology of Spatio-Temporal Chaos , 1987 .

[22]  Scott T. M. Dawson,et al.  Model Reduction for Flow Analysis and Control , 2017 .

[23]  Cosma Rohilla Shalizi.  Optimal Nonlinear Prediction of Random Fields on Networks , 2003, DMCS.

[24]  Adam Rupe,et al.  Local Causal States and Discrete Coherent Structures , 2018, Chaos.

[25]  G. Froyland,et al.  Almost-invariant sets and invariant manifolds — Connecting probabilistic and geometric descriptions of coherent structures in flows , 2009 .

[26]  George Haller,et al.  Lagrangian coherent structures: The hidden skeleton of fluid flows , 2013 .