Learning Stochastic Differential Equations With Gaussian Processes Without Gradient Matching

Cagatay Yildiz; Markus Heinonen; Jukka Intosalmi; Henrik Mannerström; Harri Lähdesmäki

doi:10.1109/MLSP.2018.8516991

Learning Stochastic Differential Equations With Gaussian Processes Without Gradient Matching

Cagatay Yildiz, Markus Heinonen, Jukka Intosalmi, Henrik Mannerström, Harri Lähdesmäki

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

24 Sitaatiot (Scopus)

Abstrakti

We introduce a novel paradigm for learning non-parametric drift and diffusion functions for stochastic differential equation (SDE). The proposed model learns to simulate path distributions that match observations with non-uniform time increments and arbitrary sparseness, which is in contrast with gradient matching that does not optimize simulated responses. We formulate sensitivity equations for learning and demonstrate that our general stochastic distribution optimisation leads to robust and efficient learning of SDE systems.

Alkuperäiskieli	Englanti
Otsikko	IEEE International Workshop on Machine Learning for Signal Processing
Kustantaja	IEEE
Sivumäärä	6
ISBN (elektroninen)	978-1-5386-5477-4
DOI - pysyväislinkit	https://doi.org/10.1109/MLSP.2018.8516991
Tila	Julkaistu - 2018
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisussa
Tapahtuma	IEEE International Workshop on Machine Learning for Signal Processing - Aalborg, Tanska Kesto: 17 syysk. 2018 → 20 syysk. 2018 Konferenssinumero: 28

Workshop

Workshop	IEEE International Workshop on Machine Learning for Signal Processing
Lyhennettä	MLSP
Maa/Alue	Tanska
Kaupunki	Aalborg
Ajanjakso	17/09/2018 → 20/09/2018

Pääsy asiakirjaan

10.1109/MLSP.2018.8516991

Muut tiedostot ja linkit

Link to publication in Scopus

Science-IT
Hakala, M. (Manager)
Perustieteiden korkeakoulu
Laitteistot/tilat: Facility

Siteeraa tätä

@inproceedings{f9f39f41cf9348e79da0e232568e032a,

title = "Learning Stochastic Differential Equations With Gaussian Processes Without Gradient Matching",

abstract = "We introduce a novel paradigm for learning non-parametric drift and diffusion functions for stochastic differential equation (SDE). The proposed model learns to simulate path distributions that match observations with non-uniform time increments and arbitrary sparseness, which is in contrast with gradient matching that does not optimize simulated responses. We formulate sensitivity equations for learning and demonstrate that our general stochastic distribution optimisation leads to robust and efficient learning of SDE systems.",

keywords = "Stochastic differential equations, Gaussian processes",

author = "Cagatay Yildiz and Markus Heinonen and Jukka Intosalmi and Henrik Mannerstr{\"o}m and Harri L{\"a}hdesm{\"a}ki",

year = "2018",

doi = "10.1109/MLSP.2018.8516991",

language = "English",

booktitle = "IEEE International Workshop on Machine Learning for Signal Processing",

publisher = "IEEE",

address = "United States",

note = "IEEE International Workshop on Machine Learning for Signal Processing, MLSP ; Conference date: 17-09-2018 Through 20-09-2018",

}

Yildiz, C, Heinonen, M, Intosalmi, J, Mannerström, H & Lähdesmäki, H 2018, Learning Stochastic Differential Equations With Gaussian Processes Without Gradient Matching. julkaisussa IEEE International Workshop on Machine Learning for Signal Processing., 8516991, IEEE, IEEE International Workshop on Machine Learning for Signal Processing, Aalborg, Tanska, 17/09/2018. https://doi.org/10.1109/MLSP.2018.8516991

Learning Stochastic Differential Equations With Gaussian Processes Without Gradient Matching. / Yildiz, Cagatay; Heinonen, Markus; Intosalmi, Jukka et al.
IEEE International Workshop on Machine Learning for Signal Processing. IEEE, 2018. 8516991.

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

TY - GEN

T1 - Learning Stochastic Differential Equations With Gaussian Processes Without Gradient Matching

AU - Yildiz, Cagatay

AU - Heinonen, Markus

AU - Intosalmi, Jukka

AU - Mannerström, Henrik

AU - Lähdesmäki, Harri

N1 - Conference code: 28

PY - 2018

Y1 - 2018

N2 - We introduce a novel paradigm for learning non-parametric drift and diffusion functions for stochastic differential equation (SDE). The proposed model learns to simulate path distributions that match observations with non-uniform time increments and arbitrary sparseness, which is in contrast with gradient matching that does not optimize simulated responses. We formulate sensitivity equations for learning and demonstrate that our general stochastic distribution optimisation leads to robust and efficient learning of SDE systems.

AB - We introduce a novel paradigm for learning non-parametric drift and diffusion functions for stochastic differential equation (SDE). The proposed model learns to simulate path distributions that match observations with non-uniform time increments and arbitrary sparseness, which is in contrast with gradient matching that does not optimize simulated responses. We formulate sensitivity equations for learning and demonstrate that our general stochastic distribution optimisation leads to robust and efficient learning of SDE systems.

KW - Stochastic differential equations

KW - Gaussian processes

UR - http://www.scopus.com/inward/record.url?scp=85055708011&partnerID=8YFLogxK

U2 - 10.1109/MLSP.2018.8516991

DO - 10.1109/MLSP.2018.8516991

M3 - Conference article in proceedings

BT - IEEE International Workshop on Machine Learning for Signal Processing

PB - IEEE

T2 - IEEE International Workshop on Machine Learning for Signal Processing

Y2 - 17 September 2018 through 20 September 2018

ER -

Learning Stochastic Differential Equations With Gaussian Processes Without Gradient Matching

Abstrakti

Workshop

Pääsy asiakirjaan

Muut tiedostot ja linkit

Sormenjälki

Laitteet

Science-IT

Siteeraa tätä