Bayes linear regression and basis-functions in Gaussian process regression

a.k.a Fixed Rank Kriging, weight space GPs

No, these are officers. You want low rank Gaussian processes.

Another way of cunningly chopping up the work of fitting a Gaussian process is to represent the process as a random function comprising basis functions $$\phi=\left(\phi_{1}, \ldots, \phi_{\ell}\right)$$ with the Gaussian random weight vector $$w$$ so that $f^{(w)}(\cdot)=\sum_{i=1}^{\ell} w_{i} \phi_{i}(\cdot) \quad \boldsymbol{w} \sim \mathcal{N}\left(\mathbf{0}, \boldsymbol{\Sigma}_{\boldsymbol{w}}\right).$ $$f^{(w)}$$ is a random function satisfying $$\boldsymbol{f}^{(\boldsymbol{w})} \sim \mathcal{N}\left(\mathbf{0}, \boldsymbol{\Phi}_{n} \boldsymbol{\Sigma}_{\boldsymbol{w}} \boldsymbol{\Phi}^{\top}\right)$$, where $$\boldsymbol{\Phi}_{n}=\boldsymbol{\phi}(\mathbf{X})$$ is a $$|\mathbf{X}| \times \ell$$ matrix of features. This is referred to as a weight space approach in ML.

TODO: I just assumed centred weights here, but that is crazy. Update to relax that assumption.

We might imagine this representation would be exact if we had countably many basis functions, and under sane conditions it is. We would like to know, further, that we can find a basis such that we need not too many basis functions to represent the process. Looking at the Karhunen-Loève theorem theorem we might imagine that this can sometimes work out fine, and indeed it does, sometimes.

This is a classic; see Chapter 3 of Bishop (2006) is classic and nicely clear. Cressie and Wikle (2011) targets for the spatiotemporal context.

Hijinks ensue when selecting the basis functions. If we were to treat the natural Hilbert space here seriously we could consider identifying the bases as eigenfunctions of the kernel. This is not generally easy. We tend to use either global bases such as Fourier bases or more generally Karhunen-Loéve bases, or construct local bases of limited overlap (usually piecewise polynomials AFAICT).

Fourier features

When the Fourier basis is natural for the problem we are in a pretty good situation. We can use the Wiener Khintchine relations to analyse and simulate the process.

Random Fourier features

The random Fourier features method constructs a Monte Carlo estimate to a stationary kernel by representing the inner product in terms of $$\ell$$ complex exponential basis functions $$\phi_{j}(\boldsymbol{x})=\ell^{-1 / 2} \exp \left(i \boldsymbol{\omega}_{j}^{\top} \boldsymbol{x}\right)$$ with frequency parameters $$\boldsymbol{\omega}_{j}$$ sampled proportionally to the spectral density $$\rho\left(\boldsymbol{\omega}_{j}\right).$$

This sometimes has a favourable error rate .

K-L basis

We recall from the Karhunen-Loéve notebook that the mean-square-optimal $$f^{(w)}$$ for approximating a Gaussian process $$f$$ is found by truncating the Karhunen-Loéve expansion $f(\cdot)=\sum_{i=1}^{\infty} w_{i} \phi_{i}(\cdot) \quad w_{i} \sim \mathcal{N}\left(0, \lambda_{i}\right)$ where $$\phi_{i}$$ and $$\lambda_{i}$$ are, respectively, the $$i$$-th (orthogonal) eigenfunction and eigenvalue of the covariance operator $$\psi \mapsto \int_{\mathcal{X}} \psi(\boldsymbol{x}) k(\boldsymbol{x}, \cdot) \mathrm{d} \boldsymbol{x}$$, written in decreasing order of $$\lambda_{i}$$. What is the orthogonal basis $$\{\phi_{i}\}_i$$ though? That depends on the problem and can be a lot of work to calculate.

In the case that our field is stationary on a “nice” domain, though, this can easy — we simply have the Fourier features as the natural basis.

Compactly-supported basis functions

As seen in GPs as SDEs and FEMs .

“Decoupled” bases

Cheng and Boots (2017); Salimbeni et al. (2018); Shi, Titsias, and Mnih (2020); Wilson et al. (2020).

References

Ambikasaran, Sivaram, Daniel Foreman-Mackey, Leslie Greengard, David W. Hogg, and Michael O’Neil. 2015. arXiv:1403.6015 [Astro-Ph, Stat], April.
Bishop, Christopher M. 2006. Pattern Recognition and Machine Learning. Information Science and Statistics. New York: Springer.
Cheng, Ching-An, and Byron Boots. 2017. In Advances in Neural Information Processing Systems. Vol. 30. Curran Associates, Inc.
Cressie, Noel, and Hsin-Cheng Huang. 1999. Journal of the American Statistical Association 94 (448): 1330–39.
Cressie, Noel, and Gardar Johannesson. 2008. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 70 (1): 209–26.
Cressie, Noel, Tao Shi, and Emily L. Kang. 2010. Journal of Computational and Graphical Statistics 19 (3): 724–45.
Cressie, Noel, and Christopher K. Wikle. 2011. Statistics for Spatio-Temporal Data. Wiley Series in Probability and Statistics 2.0. John Wiley and Sons.
———. 2014. In Wiley StatsRef: Statistics Reference Online. American Cancer Society.
Dahl, Astrid, and Edwin V. Bonilla. 2019. arXiv:1903.03986 [Cs, Stat], March.
Dubrule, Olivier. 2018. In Handbook of Mathematical Geosciences: Fifty Years of IAMG, edited by B.S. Daya Sagar, Qiuming Cheng, and Frits Agterberg, 3–24. Cham: Springer International Publishing.
Finley, Andrew O., Sudipto Banerjee, and Alan E. Gelfand. 2015. Journal of Statistical Software 63 (February): 1–28.
Ghanem, Roger, and P. D. Spanos. 1990. Journal of Applied Mechanics 57 (1): 197–202.
Gilboa, E., Y. Saatçi, and J. P. Cunningham. 2015. IEEE Transactions on Pattern Analysis and Machine Intelligence 37 (2): 424–36.
Gulian, Mamikon, Ari Frankel, and Laura Swiler. 2020. arXiv:2012.11857 [Cs, Math, Stat], December.
Hu, Xiangping, Ingelin Steinsland, Daniel Simpson, Sara Martino, and Håvard Rue. 2013. July.
Le, Quoc, Tamás Sarlós, and Alex Smola. 2013. In Proceedings of the International Conference on Machine Learning.
Lei, Huan, Jing Li, Peiyuan Gao, Panos Stinis, and Nathan Baker. 2018. April.
Lindgren, Finn, Håvard Rue, and Johan Lindström. 2011. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 73 (4): 423–98.
Liu, Chong, Surajit Ray, and Giles Hooker. 2014. arXiv:1411.4681 [Math, Stat], November.
Lord, Gabriel J., Catherine E. Powell, and Tony Shardlow. 2014. An Introduction to Computational Stochastic PDEs. 1st edition. Cambridge Texts in Applied Mathematics. New York, NY, USA: Cambridge University Press.
Luo, Wuan. 2006. Phd, California Institute of Technology.
Miller, David L., Richard Glennie, and Andrew E. Seaton. 2020. Journal of Agricultural, Biological and Environmental Statistics 25 (1): 1–16.
Nguyen, Hai, Noel Cressie, and Amy Braverman. 2012. Journal of the American Statistical Association 107 (499): 1004–18.
Nowak, W., and A. Litvinenko. 2013. Mathematical Geosciences 45 (4): 411–35.
O’Hagan, Anthony. 2013. “Polynomial Chaos: A Tutorial and Critique from a Statistician’s Perspective,” 20.
Petra, Noemi, James Martin, Georg Stadler, and Omar Ghattas. 2014. SIAM Journal on Scientific Computing 36 (4): A1525–55.
Queipo, Nestor V., Raphael T. Haftka, Wei Shyy, Tushar Goel, Rajkumar Vaidyanathan, and P. Kevin Tucker. 2005. Progress in Aerospace Sciences 41 (1): 1–28.
Rahimi, Ali, and Benjamin Recht. 2007. In Advances in Neural Information Processing Systems, 1177–84. Curran Associates, Inc.
———. 2008. In 2008 46th Annual Allerton Conference on Communication, Control, and Computing, 555–61.
———. 2009. In Advances in Neural Information Processing Systems, 1313–20. Curran Associates, Inc.
Riutort-Mayol, Gabriel, Paul-Christian Bürkner, Michael R. Andersen, Arno Solin, and Aki Vehtari. 2020. arXiv:2004.11408 [Stat], April.
Salimbeni, Hugh, Ching-An Cheng, Byron Boots, and Marc Deisenroth. 2018. In Proceedings of the 32nd International Conference on Neural Information Processing Systems, 31:8725–34. NIPS’18. Red Hook, NY, USA: Curran Associates Inc.
Särkkä, Simo, A. Solin, and J. Hartikainen. 2013. IEEE Signal Processing Magazine 30 (4): 51–61.
Shi, Jiaxin, Michalis Titsias, and Andriy Mnih. 2020. In Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, 1932–42. PMLR.
Solin, Arno, and Manon Kok. 2019. In Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics, 2193–2202. PMLR.
Stein, Michael L. 2008. Journal of the Korean Statistical Society 37 (1): 3–10.
Sutherland, Danica J., and Jeff Schneider. 2015. In Proceedings of the Thirty-First Conference on Uncertainty in Artificial Intelligence, 862–71. UAI’15. Arlington, Virginia, USA: AUAI Press.
Wilson, James T, Viacheslav Borovitskiy, Alexander Terenin, Peter Mostowsky, and Marc Deisenroth. 2020. In Proceedings of the 37th International Conference on Machine Learning, 10292–302. PMLR.
Zammit-Mangion, Andrew, and Noel Cressie. 2021. Journal of Statistical Software 98 (May): 1–48.

No comments yet. Why not leave one?

GitHub-flavored Markdown & a sane subset of HTML is supported.