Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows

George Papamakarios; David C Sterratt; Iain Murray

Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows

George Papamakarios, David C Sterratt, Iain Murray

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract / Description of output

We present Sequential Neural Likelihood (SNL), a new method for Bayesian inference in simulator models, where the likelihood is intractable but simulating data from the model is possible. SNL trains an autoregressive flow on simulated data in order to learn a model of the likelihood in the region of high posterior density. A sequential training procedure guides simulations and reduces simulation cost by orders of magnitude. We show that SNL is more robust, more accurate and requires less tuning than related neural-based methods, and we discuss diagnostics for assessing calibration, convergence and goodness-of-fit.

Original language	English
Title of host publication	Proceedings of the 22nd International Conference on Artificial Intelligence and Statistics (AISTATS) 2019
Place of Publication	Naha, Okinawa, Japan
Publisher	PMLR
Pages	837-848
Number of pages	12
Publication status	Published - 18 Apr 2019
Event	22nd International Conference on Artificial Intelligence and Statistics - Naha, Japan Duration: 16 Apr 2019 → 18 Apr 2019 https://www.aistats.org/

Publication series

Name	Proceedings of Machine Learning Research
Publisher	PMLR
Volume	89
ISSN (Electronic)	2640-3498

Conference

Conference	22nd International Conference on Artificial Intelligence and Statistics
Abbreviated title	AISTATS 2019
Country/Territory	Japan
City	Naha
Period	16/04/19 → 18/04/19
Internet address	https://www.aistats.org/

Access to Document

Sequential Neural LikelihoodAccepted author manuscript, 508 KB
Sequential Neural Likelihood_PAPAMAKARIOS_DOA22122018_VOR_CC-BYFinal published version, 508 KBLicence: Creative Commons: Attribution (CC-BY)
SNL_appendixOther version, 5.14 MB

Human Brain Project Phase 3 (SGA2)
Armstrong, D.
EU government bodies
1/04/18 → 31/03/20
Project: Research

Iain Murray
- i.murrayed.acuk
- School of Informatics - Personal Chair of Machine Learning and Inference
- Institute for Adaptive and Neural Computation
- Data Science and Artificial Intelligence
Person: Academic: Research Active

Cite this

Papamakarios, G., Sterratt, D. C., & Murray, I. (2019). Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows. In Proceedings of the 22nd International Conference on Artificial Intelligence and Statistics (AISTATS) 2019 (pp. 837-848). (Proceedings of Machine Learning Research; Vol. 89). PMLR. http://proceedings.mlr.press/v89/papamakarios19a.html

@inproceedings{99aad490fd134b849b5336473b4342c3,

title = "Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows",

abstract = "We present Sequential Neural Likelihood (SNL), a new method for Bayesian inference in simulator models, where the likelihood is intractable but simulating data from the model is possible. SNL trains an autoregressive flow on simulated data in order to learn a model of the likelihood in the region of high posterior density. A sequential training procedure guides simulations and reduces simulation cost by orders of magnitude. We show that SNL is more robust, more accurate and requires less tuning than related neural-based methods, and we discuss diagnostics for assessing calibration, convergence and goodness-of-fit.",

author = "George Papamakarios and Sterratt, {David C} and Iain Murray",

year = "2019",

month = apr,

day = "18",

language = "English",

series = "Proceedings of Machine Learning Research",

publisher = "PMLR",

pages = "837--848",

booktitle = "Proceedings of the 22nd International Conference on Artificial Intelligence and Statistics (AISTATS) 2019",

note = "22nd International Conference on Artificial Intelligence and Statistics, AISTATS 2019 ; Conference date: 16-04-2019 Through 18-04-2019",

url = "https://www.aistats.org/",

}

Papamakarios, G, Sterratt, DC & Murray, I 2019, Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows. in Proceedings of the 22nd International Conference on Artificial Intelligence and Statistics (AISTATS) 2019. Proceedings of Machine Learning Research, vol. 89, PMLR, Naha, Okinawa, Japan, pp. 837-848, 22nd International Conference on Artificial Intelligence and Statistics, Naha, Japan, 16/04/19. <http://proceedings.mlr.press/v89/papamakarios19a.html>

Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows. / Papamakarios, George; Sterratt, David C ; Murray, Iain.
Proceedings of the 22nd International Conference on Artificial Intelligence and Statistics (AISTATS) 2019. Naha, Okinawa, Japan: PMLR, 2019. p. 837-848 (Proceedings of Machine Learning Research; Vol. 89).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows

AU - Papamakarios, George

AU - Sterratt, David C

AU - Murray, Iain

PY - 2019/4/18

Y1 - 2019/4/18

N2 - We present Sequential Neural Likelihood (SNL), a new method for Bayesian inference in simulator models, where the likelihood is intractable but simulating data from the model is possible. SNL trains an autoregressive flow on simulated data in order to learn a model of the likelihood in the region of high posterior density. A sequential training procedure guides simulations and reduces simulation cost by orders of magnitude. We show that SNL is more robust, more accurate and requires less tuning than related neural-based methods, and we discuss diagnostics for assessing calibration, convergence and goodness-of-fit.

AB - We present Sequential Neural Likelihood (SNL), a new method for Bayesian inference in simulator models, where the likelihood is intractable but simulating data from the model is possible. SNL trains an autoregressive flow on simulated data in order to learn a model of the likelihood in the region of high posterior density. A sequential training procedure guides simulations and reduces simulation cost by orders of magnitude. We show that SNL is more robust, more accurate and requires less tuning than related neural-based methods, and we discuss diagnostics for assessing calibration, convergence and goodness-of-fit.

M3 - Conference contribution

T3 - Proceedings of Machine Learning Research

SP - 837

EP - 848

BT - Proceedings of the 22nd International Conference on Artificial Intelligence and Statistics (AISTATS) 2019

PB - PMLR

CY - Naha, Okinawa, Japan

T2 - 22nd International Conference on Artificial Intelligence and Statistics

Y2 - 16 April 2019 through 18 April 2019

ER -

Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows

Abstract / Description of output

Publication series

Conference

Access to Document

Fingerprint

Projects

Human Brain Project Phase 3 (SGA2)

Profiles

Iain Murray

Cite this