Learning stable and predictive structures in kinetic systems

Publikation: Bidrag til tidsskriftTidsskriftartikelfagfællebedømt

Standard

Learning stable and predictive structures in kinetic systems. / Pfister, Niklas; Bauer, Stefan; Peters, Jonas.

I: Proceedings of the National Academy of Sciences of the United States of America, Bind 116, Nr. 51, 2019, s. 25405-25411.

Publikation: Bidrag til tidsskriftTidsskriftartikelfagfællebedømt

Harvard

Pfister, N, Bauer, S & Peters, J 2019, 'Learning stable and predictive structures in kinetic systems', Proceedings of the National Academy of Sciences of the United States of America, bind 116, nr. 51, s. 25405-25411. https://doi.org/10.1073/pnas.1905688116

APA

Pfister, N., Bauer, S., & Peters, J. (2019). Learning stable and predictive structures in kinetic systems. Proceedings of the National Academy of Sciences of the United States of America, 116(51), 25405-25411. https://doi.org/10.1073/pnas.1905688116

Vancouver

Pfister N, Bauer S, Peters J. Learning stable and predictive structures in kinetic systems. Proceedings of the National Academy of Sciences of the United States of America. 2019;116(51):25405-25411. https://doi.org/10.1073/pnas.1905688116

Author

Pfister, Niklas ; Bauer, Stefan ; Peters, Jonas. / Learning stable and predictive structures in kinetic systems. I: Proceedings of the National Academy of Sciences of the United States of America. 2019 ; Bind 116, Nr. 51. s. 25405-25411.

Bibtex

@article{aa20dde4a6f640119cdf954a83df2dc3,
title = "Learning stable and predictive structures in kinetic systems",
abstract = "Learning kinetic systems from data is one of the core challenges in many fields. Identifying stable models is essential for the generalization capabilities of data-driven inference. We introduce a computationally efficient framework, called CausalKinetiX, that identifies structure from discrete time, noisy observations, generated from heterogeneous experiments. The algorithm assumes the existence of an underlying, invariant kinetic model, a key criterion for reproducible research. Results on both simulated and real-world examples suggest that learning the structure of kinetic systems benefits from a causal perspective. The identified variables and models allow for a concise description of the dynamics across multiple experimental settings and can be used for prediction in unseen experiments. We observe significant improvements compared to well-established approaches focusing solely on predictive performance, especially for out-of-sample generalization",
keywords = "kinetic systems, causal inference, stability, invariance, structure learning",
author = "Niklas Pfister and Stefan Bauer and Jonas Peters",
year = "2019",
doi = "10.1073/pnas.1905688116",
language = "English",
volume = "116",
pages = "25405--25411",
journal = "Proceedings of the National Academy of Sciences of the United States of America",
issn = "0027-8424",
publisher = "The National Academy of Sciences of the United States of America",
number = "51",

}

RIS

TY - JOUR

T1 - Learning stable and predictive structures in kinetic systems

AU - Pfister, Niklas

AU - Bauer, Stefan

AU - Peters, Jonas

PY - 2019

Y1 - 2019

N2 - Learning kinetic systems from data is one of the core challenges in many fields. Identifying stable models is essential for the generalization capabilities of data-driven inference. We introduce a computationally efficient framework, called CausalKinetiX, that identifies structure from discrete time, noisy observations, generated from heterogeneous experiments. The algorithm assumes the existence of an underlying, invariant kinetic model, a key criterion for reproducible research. Results on both simulated and real-world examples suggest that learning the structure of kinetic systems benefits from a causal perspective. The identified variables and models allow for a concise description of the dynamics across multiple experimental settings and can be used for prediction in unseen experiments. We observe significant improvements compared to well-established approaches focusing solely on predictive performance, especially for out-of-sample generalization

AB - Learning kinetic systems from data is one of the core challenges in many fields. Identifying stable models is essential for the generalization capabilities of data-driven inference. We introduce a computationally efficient framework, called CausalKinetiX, that identifies structure from discrete time, noisy observations, generated from heterogeneous experiments. The algorithm assumes the existence of an underlying, invariant kinetic model, a key criterion for reproducible research. Results on both simulated and real-world examples suggest that learning the structure of kinetic systems benefits from a causal perspective. The identified variables and models allow for a concise description of the dynamics across multiple experimental settings and can be used for prediction in unseen experiments. We observe significant improvements compared to well-established approaches focusing solely on predictive performance, especially for out-of-sample generalization

KW - kinetic systems

KW - causal inference

KW - stability

KW - invariance

KW - structure learning

U2 - 10.1073/pnas.1905688116

DO - 10.1073/pnas.1905688116

M3 - Journal article

C2 - 31776252

VL - 116

SP - 25405

EP - 25411

JO - Proceedings of the National Academy of Sciences of the United States of America

JF - Proceedings of the National Academy of Sciences of the United States of America

SN - 0027-8424

IS - 51

ER -

ID: 233585851