Robust Regression for Automatic Fusion Plasma Analysis Based on Generative Modeling

The first step to realize an automatic experimental data analysis for fusion plasma experiments is fitting noisy data of temperature and density spatial profiles, which are routinely obtained. However, it has been difficult to construct algorithms that fit all the data without overfitting and underfitting. In this paper, we show that this difficulty originates from the lack of knowledge of the probability distribution that the measurement data follow. We demonstrate the use of a machine learning technique to estimate the data distribution and to construct an optimal generative model. We show that the fitting algorithm based on the generative modeling outperforms classical heuristic methods in terms of the stability as well as the accuracy.

[1]  Manfred Opper,et al.  The Variational Gaussian Approximation Revisited , 2009, Neural Computation.

[2]  W. Cleveland.  Robust Locally Weighted Regression and Smoothing Scatterplots , 1979 .

[3]  D. F. Andrews,et al.  A Robust Method for Multiple Linear Regression , 1974 .

[4]  K. Fujii,et al.  Machine Learning of Noise in LHD Thomson Scattering System , 2018 .

[5]  W. Härdle,et al.  Robust Non-parametric Function Fitting , 1984 .

[6]  V. Altuzar,et al.  Atmospheric pollution profiles in Mexico City in two different seasons , 2003 .

[7]  Keisuke Fujii,et al.  Data-driven sensitivity inference for Thomson scattering electron density measurement systems. , 2016, The Review of scientific instruments.

[8]  Chihiro Suzuki,et al.  Development and application of real-time magnetic coordinate mapping system in the Large Helical Device , 2012 .

[9]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[10]  Akio Sanpei,et al.  Extension of the operational regime of the LHD towards a deuterium experiment , 2017 .

[11]  D. Brower,et al.  Advanced plasma fluctuation analysis techniques and their impact on fusion research (invited) , 1988 .

[12]  I. Yamada,et al.  Raman calibration of the LHD YAG Thomson scattering for electron-density measurements , 2003 .

[13]  Michael I. Jordan,et al.  An Introduction to Variational Methods for Graphical Models , 1999, Machine Learning.

[14]  O. William.  Journal Of The American Statistical Association V-28 , 1932 .

[15]  Shakir Mohamed,et al.  Variational Inference with Normalizing Flows , 2015, ICML.

[16]  Youssef M. Marzouk,et al.  Improved profile fitting and quantification of uncertainty in experimental measurements of impurity transport coefficients using Gaussian process regression , 2015 .

[17]  I. Yamada,et al.  Recent Progress of the LHD Thomson Scattering System , 2010 .

[18]  Harvey J. Motulsky,et al.  Detecting outliers when fitting data with nonlinear regression – a new method based on robust nonlinear regression and the false discovery rate , 2006, BMC Bioinformatics.

[19]  D. A. Gates,et al.  Extended capability of the integrated transport analysis suite, TASK3D-a, for LHD experiment , 2017 .