Path Signatures on Lie Groups

Path signatures are powerful nonparametric tools for time series analysis, shown to form a universal and characteristic feature map for Euclidean valued time series data. We lift the theory of path signatures to the setting of Lie group valued time series, adapting these tools for time series with underlying geometric constraints. We prove that this generalized path signature is universal and characteristic. To demonstrate universality, we analyze the human action recognition problem in computer vision, using $SO(3)$ representations for the time series, providing comparable performance to other shallow learning approaches, while offering an easily interpretable feature set. We also provide a two-sample hypothesis test for Lie group-valued random walks to illustrate its characteristic property. Finally we provide algorithms and a Julia implementation of these methods.

[1]  Dimitrios Makris,et al.  G3D: A gaming action dataset and real time action recognition evaluation framework , 2012, 2012 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops.

[2]  Elena Celledoni,et al.  Signatures in Shape Analysis: an Efficient Approach to Motion Identification , 2019, GSI.

[3]  Peter K. Friz,et al.  Multidimensional Stochastic Processes as Rough Paths: Theory and Applications , 2010 .

[4]  Terry Lyons,et al.  Uniqueness for the signature of a path of bounded variation and the reduced path group , 2005, math/0507536.

[5]  P. Diaconis,et al.  The Subgroup Algorithm for Generating Uniform Random Variables , 1987, Probability in the Engineering and Informational Sciences.

[6]  Bernd Sturmfels,et al.  Learning Paths from Signature Tensors , 2018, SIAM J. Matrix Anal. Appl..

[7]  I. Chevyrev,et al.  Signature Moments to Characterize Laws of Stochastic Processes , 2018, J. Mach. Learn. Res..

[8]  E. Donne.  A Primer on Carnot Groups: Homogenous Groups, Carnot-Carathéodory Spaces, and Regularity of Their Isometries , 2018 .

[9]  Terry Lyons,et al.  Hyperbolic development and inversion of signature , 2015, 1507.00286.

[10]  Terry Lyons,et al.  A feature set for streams and an application to high-frequency financial tick data , 2014, BigDataScience '14.

[11]  Yanshan Li,et al.  Skeleton-based Action Recognition with Lie Group and Deep Neural Networks , 2019, 2019 IEEE 4th International Conference on Signal and Image Processing (ICSIP).

[12]  Lianwen Jin,et al.  DeepWriterID: An End-to-End Online Text-Independent Writer Identification System , 2015, IEEE Intelligent Systems.

[13]  Rama Chellappa,et al.  Human Action Recognition by Representing 3D Skeletons as Points in a Lie Group , 2014, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[14]  Kuo-Tsai Chen,et al.  Iterated Integrals and Exponential Homomorphisms , 1954 .

[15]  Chad Giusti,et al.  Iterated Integrals and Population Time Series Analysis , 2020 .

[16]  Benjamin Graham,et al.  The iisignature library: efficient calculation of iterated-integral signatures and log signatures , 2017, ACM Trans. Math. Softw..

[17]  Cordelia Schmid,et al.  Leveraging the Path Signature for Skeleton-based Human Action Recognition , 2017, ArXiv.

[18]  Benjamin J. Zimmerman,et al.  Dissociating tinnitus patients from healthy controls using resting-state cyclicity analysis and clustering , 2018, Network Neuroscience.

[19]  R. Giles.  A generalization of the strict topology , 1971 .

[20]  Kuo-Tsai Ciren,et al.  INTEGRATION OF PATHS, GEOMETRIC INVARIANTS AND A GENERALIZED BAKER-HAUSDORFF FORMULA , 2016 .

[21]  Andrey Kormilitzin,et al.  A Primer on the Signature Method in Machine Learning , 2016, ArXiv.

[22]  Yuliy Baryshnikov,et al.  Cyclicity in multivariate time series and applications to functional MRI data , 2016, 2016 IEEE 55th Conference on Decision and Control (CDC).

[23]  Bernhard Schölkopf,et al.  Hilbert Space Embeddings and Metrics on Probability Measures , 2009, J. Mach. Learn. Res..

[24]  Kuo-Tsai Chen.  INTEGRATION OF PATHS—A FAITHFUL REPRE- SENTATION OF PATHS BY NONCOMMUTATIVE FORMAL POWER SERIES , 1958 .

[25]  Terry Lyons,et al.  Extracting information from the signature of a financial data stream , 2013, 1307.7244.

[26]  Terry Lyons,et al.  A signature-based machine learning model for distinguishing bipolar disorder and borderline personality disorder , 2017, Translational Psychiatry.

[27]  Zaïd Harchaoui,et al.  A Fast, Consistent Kernel Two-Sample Test , 2009, NIPS.

[28]  Kuo-Tsai Chen,et al.  Iterated path integrals , 1977 .

[29]  Signature inversion for monotone paths , 2016, 1612.04454.

[30]  Imed Riadh Farah,et al.  Action Recognition from 3D Skeleton Sequences using Deep Networks on Lie Group Features , 2018, 2018 24th International Conference on Pattern Recognition (ICPR).

[31]  Kurusch Ebrahimi-Fard,et al.  Time-Warping Invariants of Multidimensional Time Series , 2019, ArXiv.

[32]  Annegret Y. Burtscher.  Length structures on manifolds with continuous Riemannian metrics , 2012, 1212.6962.

[33]  Luc Van Gool,et al.  Deep Learning on Lie Groups for Skeleton-Based Action Recognition , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[34]  J. M. Selig.  Lie Groups and Lie Algebras in Robotics , 2004 .

[35]  Hans-Peter Kriegel,et al.  Integrating structured biological data by Kernel Maximum Mean Discrepancy , 2006, ISMB.

[36]  Adeline Fermanian.  Embedding and learning with signatures , 2019, ArXiv.

[37]  École d'été de probabilités de Saint-Flour,et al.  Differential equations driven by rough paths , 2007 .

[38]  Gérard Dreyfus,et al.  Single-layer learning revisited: a stepwise procedure for building and training a neural network , 1989, NATO Neurocomputing.

[39]  Terry Lyons.  Rough paths, Signatures and the modelling of functions on streams , 2014, 1405.4537.

[40]  John M. Lee.  Introduction to Smooth Manifolds , 2002 .

[41]  B. Sturmfels,et al.  VARIETIES OF SIGNATURE TENSORS , 2018, Forum of Mathematics, Sigma.

[42]  S. Strogatz.  From Kuramoto to Crawford: exploring the onset of synchronization in populations of coupled oscillators , 2000 .

[43]  Harald Oberhauser,et al.  Persistence Paths and Signature Features in Topological Data Analysis , 2018, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[44]  P J Moore,et al.  Using path signatures to predict a diagnosis of Alzheimer’s disease , 2018, PloS one.

[45]  Terry Lyons,et al.  System Control and Rough Paths , 2003 .

[46]  Bernhard Schölkopf,et al.  Kernel Distribution Embeddings: Universal Kernels, Characteristic Kernels and Kernel Metrics on Distributions , 2016, J. Mach. Learn. Res..

[47]  Terry Lyons.  Di erential equations driven by rough signals , 1998 .

[48]  Patrick Kidger,et al.  Signatory: differentiable computations of the signature and logsignature transforms, on both CPU and GPU , 2020, ArXiv.

[49]  Franz J. Király,et al.  Kernels for sequentially ordered data , 2016, J. Mach. Learn. Res..

[50]  Bernhard Schölkopf,et al.  A Kernel Two-Sample Test , 2012, J. Mach. Learn. Res..

[51]  Rama Chellappa,et al.  Rolling Rotations for Recognizing Human Actions from 3D Skeletal Data , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[52]  Philipp Probst,et al.  Hyperparameters and tuning strategies for random forest , 2018, WIREs Data Mining Knowl. Discov..

[53]  Joscha Diehl,et al.  Invariants of Multidimensional Time Series Based on Their Iterated-Integral Signature , 2018, Acta Applicandae Mathematicae.

[54]  Marcos M. Alexandrino,et al.  Lie Groups and Geometric Aspects of Isometric Actions , 2015 .