Identification of Anomalous Diffusion Sources by Unsupervised Learning

Fractional Brownian motion (fBm) is a ubiquitous diffusion process in which the memory effects of the stochastic transport result in the mean squared particle displacement following a power law, $\langle {\Delta r}^2 \rangle \sim t^{\alpha}$, where the diffusion exponent $\alpha$ characterizes whether the transport is subdiffusive, ($\alpha 1$). Due to the abundance of fBm processes in nature, significant efforts have been devoted to the identification and characterization of fBm sources in various phenomena. In practice, the identification of the fBm sources often relies on solving a complex and ill-posed inverse problem based on limited observed data. In the general case, the detected signals are formed by an unknown number of release sources, located at different locations and with different strengths, that act simultaneously. This means that the observed data is composed of mixtures of releases from an unknown number of sources, which makes the traditional inverse modeling approaches unreliable. Here, we report an unsupervised learning method, based on Nonnegative Matrix Factorization, that enables the identification of the unknown number of release sources as well the anomalous diffusion characteristics based on limited observed data and the general form of the corresponding fBm Green's function. We show that our method performs accurately for different types of sources and configurations with a predetermined number of sources with specific characteristics and introduced noise.

[1]  M. Stratton,et al.  Deciphering Signatures of Mutational Processes Operative in Human Cancer , 2013, Cell reports.

[2]  J. Janno,et al.  Reconstruction of an order of derivative and a source term in a fractional diffusion equation from final measurements , 2018 .

[3]  Zhang Weiguo,et al.  Exact maximum likelihood estimators for drift fractional Brownian motions , 2009, 0904.4186.

[4]  Erik Skau,et al.  Distributed non-negative matrix factorization with determination of the number of latent features , 2020, The Journal of Supercomputing.

[5]  M. Berberan-Santos,et al.  Photon trajectories in incoherent atomic radiation trapping as Lévy flights. , 2003, Physical review letters.

[6]  H. Sebastian Seung,et al.  Learning the parts of objects by non-negative matrix factorization , 1999, Nature.

[7]  Kevin E. Bassler,et al.  Hurst exponents, Markov processes, and fractional Brownian motion , 2006, cond-mat/0609671.

[8]  Wen Chen,et al.  A coupled method for inverse source problem of spatial fractional anomalous diffusion equations , 2010 .

[9]  Yuan-Xiang Zhang,et al.  Bayesian approach to a nonlinear inverse problem for a time-space fractional diffusion equation , 2018, Inverse Problems.

[10]  J. Klafter,et al.  The random walk's guide to anomalous diffusion: a fractional dynamics approach , 2000 .

[11]  B. Berkowitz,et al.  Anomalous Transport in “Classical” Soil and Sand Columns , 2004, Soil Science Society of America Journal.

[12]  P. Rousseeuw Silhouettes: a graphical aid to the interpretation and validation of cluster analysis , 1987 .

[13]  A. Caspi,et al.  Enhanced diffusion in active intracellular transport. , 2000, Physical review letters.

[14]  Weilin Xiao,et al.  Parameter identification for mixed fractional Brownian motions with the drift parameter , 2019 .

[15]  Zhi-Hua Zhou,et al.  Non-negative matrix factorization on Kernels , 2006 .

[16]  J. Klafter,et al.  The restaurant at the end of the random walk: recent developments in the description of anomalous transport by fractional dynamics , 2004 .

[17]  D. Rubin,et al.  Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper , 1977 .

[18]  B. Øksendal,et al.  Stochastic Calculus for Fractional Brownian Motion and Applications , 2008 .

[19]  Daoqiang Zhang,et al.  Non-negative Matrix Factorization on Kernels , 2006, PRICAI.

[20]  G. Kneller,et al.  Communication: consistent picture of lateral subdiffusion in lipid bilayers: molecular dynamics simulation and exact results. , 2011, The Journal of chemical physics.

[21]  Carl-Fredrik Westin,et al.  Estimating Diffusion Propagator and Its Moments Using Directional Radial Basis Functions , 2015, IEEE Transactions on Medical Imaging.

[22]  J. Coeurjolly,et al.  Estimating the Parameters of a Fractional Brownian Motion by Discrete Variations of its Sample Paths , 2001 .

[23]  Masahiro Yamamoto,et al.  Uniqueness in an inverse problem for a one-dimensional fractional diffusion equation , 2009 .

[24]  David T. W. Jones,et al.  Signatures of mutational processes in human cancer , 2013, Nature.

[25]  J. Bouchaud,et al.  Anomalous diffusion in disordered media: Statistical mechanisms, models and physical applications , 1990 .

[26]  Fred J. Molz,et al.  Fractional Brownian motion and fractional Gaussian noise in subsurface hydrology: A review, presentation of fundamental properties, and extensions , 1997 .

[27]  I. Jolliffe Principal Component Analysis , 2005 .

[28]  G. K. Boman,et al.  A fractal‐based stochastic interpolation scheme in subsurface hydrology , 1993 .

[29]  Andrzej Cichocki,et al.  Nonnegative Matrix and Tensor Factorization T , 2007 .

[30]  B. Mandelbrot,et al.  Fractional Brownian Motions, Fractional Noises and Applications , 1968 .

[31]  Ichiro Takeuchi,et al.  Unsupervised phase mapping of X-ray diffraction data by nonnegative matrix factorization integrated with custom clustering , 2018, npj Computational Materials.

[32]  M. Modugno,et al.  Observation of subdiffusion in a disordered interacting system. , 2010, Physical review letters.

[33]  William P. Ball,et al.  Application of inverse methods to contaminant source identification from aquitard diffusion profiles at Dover AFB, Delaware , 1999 .

[34]  Khaled M. Furati,et al.  An inverse problem for a generalized fractional diffusion , 2014, Appl. Math. Comput..

[35]  Q. Chen,et al.  Identification of contaminant sources in enclosed environments by inverse CFD modeling. , 2007, Indoor air.

[36]  Velimir V. Vesselinov,et al.  Identification of release sources in advection–diffusion system by machine learning combined with Green’s function inverse method , 2018, Applied Mathematical Modelling.

[37]  Martti Juhola,et al.  On principal component analysis, cosine and Euclidean measures in information retrieval , 2007, Inf. Sci..

[38]  R. Metzler,et al.  Superdiffusion dominates intracellular particle motion in the supercrowded cytoplasm of pathogenic Acanthamoeba castellanii , 2015, Scientific Reports.

[39]  Kestutis Staliunas,et al.  Anticorrelations and subdiffusion in Financial Systems , 2002, Adv. Complex Syst..

[40]  Velimir V. Vesselinov,et al.  Blind source separation for groundwater pressure analysis based on nonnegative matrix factorization , 2014 .

[41]  M. Modugno,et al.  Subdiffusion of nonlinear waves in quasiperiodic potentials , 2012, 1206.0833.

[42]  William Rundell,et al.  A tutorial on inverse problems for anomalous diffusion processes , 2015, 1501.00251.

[43]  Golan Bel,et al.  Weak Ergodicity Breaking in the Continuous-Time Random Walk , 2005 .

[44]  Yuan-Chung Cheng,et al.  Aging continuous time random walks , 2003 .

[45]  Ali Taylan Cemgil,et al.  Nonnegative matrix factorizations as probabilistic inference in composite models , 2009, 2009 17th European Signal Processing Conference.