Learning Optimal Test Statistics in the Presence of Nuisance Parameters - INSPIRE

DataBETA

Learning Optimal Test Statistics in the Presence of Nuisance Parameters

Lukas Heinrich

Mar 24, 2022

e-Print:

2203.13079 [stat.ME]

View in:

ADS Abstract Service

reference search6 citations

Citations per year

Abstract: (submitter)

The design of optimal test statistics is a key task in frequentist statistics and for a number of scenarios optimal test statistics such as the profile-likelihood ratio are known. By turning this argument around we can find the profile likelihood ratio even in likelihood-free cases, where only samples from a simulator are available, by optimizing a test statistic within those scenarios. We propose a likelihood-free training algorithm that produces test statistics that are equivalent to the profile likelihood ratios in cases where the latter is known to be optimal.

References(21)

Figures(2)

[1]

The frontier of simulation-based inference

Kyle Cranmer
(
- New York U., CCPP
)
,
Johann Brehmer
(
- New York U., CCPP
)
,
Gilles Louppe
(
- Liege U.
)

- Proc.Nat.Acad.Sci. 117 (2020) 48, 30055-30062
•
e-Print:
- 1911.01429
•
DOI:
- 10.1073/pnas.1912789117

[1]

The frontier of simulation-based inference

Kyle Cranmer
(
- New York U., CCPP
)
,
Johann Brehmer
(
- New York U., CCPP
)
,
Gilles Louppe
(
- Liege U.
)

- Proc.Nat.Acad.Sci. 117 (2020) 48, 30055-30062
•
e-Print:
- 1911.01429
•
DOI:
- 10.1073/pnas.1912789117

[2]

CoRR abs/

Tejero-Cantero
,
Boelts J
,
Deistler M
,
Lueckmann J
,
Durkan C

et al.

•
e-Print:
- 2007.09114

[2]

(Preprint

Tejero-Cantero
,
Boelts J
,
Deistler M
,
Lueckmann J
,
Durkan C

et al.

e-Print:
- 2007.09114
•
- https://arxiv.org/abs/2007.09114

[3]

Approximating Likelihood Ratios with Calibrated Discriminative Classifiers

Kyle Cranmer
(
- New York U.
)
,
Juan Pavez
(
- Santa Maria U., Valparaiso
)
,
Gilles Louppe
(
- New York U.
)

e-Print:
- 1506.02169

[4]

URL

Dalmasso N
,
Izbicki R
,
Lee A. B

•
e-Print:
- 2002.10399

[5]

Uncertainty-aware machine learning for high energy physics

Aishik Ghosh
(
- UC, Irvine and
- LBL, Berkeley
)
,
Benjamin Nachman
(
- LBL, Berkeley and
- UC, Berkeley, Miller Inst.
)
,
Daniel Whiteson
(
- UC, Irvine
)

- Phys.Rev.D 104 (2021) 5, 056026
•
e-Print:
- 2105.08742
•
DOI:
- 10.1103/PhysRevD.104.056026

[6]

Differentiable Matrix Elements with MadJax

Lukas Heinrich
(
- Munich, Tech. U.
)
,
Michael Kagan
(
- SLAC
)

- J.Phys.Conf.Ser. 2438 (2023) 1, 012137
•
e-Print:
- 2203.00057
•
DOI:
- 10.1088/1742-6596/2438/1/012137

[7]

INFERNO: Inference-Aware Neural Optimisation

Pablo De Castro
(
- INFN, Padua
)
,
Tommaso Dorigo
(
- INFN, Padua
)

- Comput.Phys.Commun. 244 (2019) 170-179
•
e-Print:
- 1806.04743
•
DOI:
- 10.1016/j.cpc.2019.06.007

[8]

neos: End-to-End-Optimised Summary Statistics for High Energy Physics

Nathan Simpson
(
- Lund U.
)
,
Lukas Heinrich
(
- Munich, Tech. U.
)

- J.Phys.Conf.Ser. 2438 (2023) 1, 012105
•
e-Print:
- 2203.05570
•
DOI:
- 10.1088/1742-6596/2438/1/012105

[9]

Philosophical Transactions of the Royal Society of London. Series A

Neyman J
,
Pearson E. S

[10]

ISSN 00029947 URL

Wald A

- Trans.Am.Math.Soc. 54 (1943) 426-482
•
- http://www.jstor.org/stable/1990256

[11]

Adam: A Method for Stochastic Optimization

Diederik P. Kingma
(
- Amsterdam U.
)
,
Jimmy Ba
(
- Toronto U.
)

- Published as a conference paper at the 3rd International Conference for Learning Representations, San Diego, 2015
•
e-Print:
- 1412.6980

[12]

Advances in Neural Information Processing Systems 32 Wallach H, Larochelle H, Beygelzimer A, d'Alché-Buc F, Fox E and Garnett R (eds.)

Paszke A
,
Gross S
,
Massa F
,
Lerer A
,
Bradbury J

et al.

[13]

Scikit-learn: Machine Learning in Python

Fabian Pedregosa
(
- INRIA, Saclay
)
,
Gaël Varoquaux
(
- INRIA, Saclay
)
,
Alexandre Gramfort
(
- INRIA, Saclay
)
,
Olivier Grisel
(
- Unlisted, FR
)
,
Mathieu Blondel
(
- Kobe U.
)

et al.

- J.Machine Learning Res. 12 (2011) 2825-2830
•
e-Print:
- 1201.0490

[14]

Asymptotic formulae for likelihood-based tests of new physics

Glen Cowan
(
- Royal Holloway, U. of London
)
,
Kyle Cranmer
(
- New York U.
)
,
Eilam Gross
(
- Weizmann Inst.
)
,
Ofer Vitells
(
- Weizmann Inst.
)

- Eur.Phys.J.C 71 (2011) 1554,
- Eur.Phys.J.C 73 (2013) 2501 (erratum)
•
e-Print:
- 1007.1727
•
DOI:
- 10.1140/epjc/s10052-011-1554-0

[14]

Asymptotic formulae for likelihood-based tests of new physics

Glen Cowan
(
- Royal Holloway, U. of London
)
,
Kyle Cranmer
(
- New York U.
)
,
Eilam Gross
(
- Weizmann Inst.
)
,
Ofer Vitells
(
- Weizmann Inst.
)

- Eur.Phys.J.C 71 (2011) 1554,
- Eur.Phys.J.C 73 (2013) 2501 (erratum)
•
e-Print:
- 1007.1727
•
DOI:
- 10.1140/epjc/s10052-011-1554-0

[15]

Array programming with NumPy

Charles R. Harris
(
- Unlisted, US, UT
)
,
K. Jarrod Millman
(
- UC, Berkeley and
- LBL, Berkeley
)
,
Stéfan J. van der Walt
(
- Stellenbosch U. and
- UC, Berkeley
)
,
Ralf Gommers
(
- Texas U.
)
,
Pauli Virtanen
(
- Jyvaskyla U.
)

et al.

- Nature 585 (2020) 7825, 357-362
•
e-Print:
- 2006.10256
•
DOI:
- 10.1038/s41586-020-2649-2

[16]

SciPy 1.0--Fundamental Algorithms for Scientific Computing in Python

Pauli Virtanen
(
- Jyvaskyla U.
)
,
Ralf Gommers
(
- Texas U.
)
,
Travis E. Oliphant
(
- Mayo Clinic and
- Brigham Young U. and
- SLAC and
- Texas U.
)
,
Matt Haberland
(
- Cal. Poly. and
- UCLA
)
,
Tyler Reddy
(
- Los Alamos
)

et al.

- Nature Meth. 17 (2020) 261
•
e-Print:
- 1907.10121
•
DOI:
- 10.1038/s41592-019-0686-2

[17]

Matplotlib: A 2D Graphics Environment

John D. Hunter
(
- Utah U.
)

- Comput.Sci.Eng. 9 (2007) 3, 90-95
•
DOI:
- 10.1109/MCSE.2007.55

[18]

Positioning and Power in

Kluyver T
,
Ragan-Kelley B
,
Pérez F
,
Granger B
,
Bussonnier M

et al.