Show simple item record

dc.contributor.author
Herrera, Calypso
dc.contributor.author
Krach, Florian
dc.contributor.author
Teichmann, Josef
dc.date.accessioned
2021-11-18T08:34:10Z
dc.date.available
2021-11-17T08:49:32Z
dc.date.available
2021-11-18T08:34:10Z
dc.date.issued
2021
dc.identifier.uri
http://hdl.handle.net/20.500.11850/515527
dc.description.abstract
Combinations of neural ODEs with recurrent neural networks (RNN), like GRU-ODE-Bayes or ODE-RNN are well suited to model irregularly observed time series. While those models outperform existing discrete-time approaches, no theoretical guarantees for their predictive capabilities are available. Assuming that the irregularly-sampled time series data originates from a continuous stochastic process, the L2-optimal online prediction is the conditional expectation given the currently available information. We introduce the Neural Jump ODE (NJ-ODE) that provides a data-driven approach to learn, continuously in time, the conditional expectation of a stochastic process. Our approach models the conditional expectation between two observations with a neural ODE and jumps whenever a new observation is made. We define a novel training framework, which allows us to prove theoretical guarantees for the first time. In particular, we show that the output of our model converges to the L2-optimal prediction. This can be interpreted as solution to a special filtering problem. We provide experiments showing that the theoretical results also hold empirically. Moreover, we experimentally show that our model outperforms the baselines in more complex learning tasks and give comparisons on real-world datasets.
en_US
dc.language.iso
en
en_US
dc.publisher
OpenReview
en_US
dc.subject
Neural ODE
en_US
dc.subject
conditional expectation
en_US
dc.subject
irregular-observed data modelling
en_US
dc.title
Neural Jump Ordinary Differential Equations: Consistent Continuous-Time Prediction and Filtering
en_US
dc.type
Conference Paper
ethz.book.title
International Conference on Learning Representations (ICLR 2021)
en_US
ethz.size
44 p.
en_US
ethz.event
9th International Conference on Learning Representations (ICLR 2021)
en_US
ethz.event.location
Online
en_US
ethz.event.date
May 3-7, 2021
en_US
ethz.publication.place
s.l.
en_US
ethz.publication.status
published
en_US
ethz.leitzahl
ETH Zürich::00002 - ETH Zürich::00012 - Lehre und Forschung::00007 - Departemente::02000 - Dep. Mathematik / Dep. of Mathematics
en_US
ethz.leitzahl
ETH Zürich::00002 - ETH Zürich::00012 - Lehre und Forschung::00007 - Departemente::02000 - Dep. Mathematik / Dep. of Mathematics::02003 - Mathematik Selbständige Professuren::03845 - Teichmann, Josef / Teichmann, Josef
en_US
ethz.leitzahl.certified
ETH Zürich::00002 - ETH Zürich::00012 - Lehre und Forschung::00007 - Departemente::02000 - Dep. Mathematik / Dep. of Mathematics::02003 - Mathematik Selbständige Professuren::03845 - Teichmann, Josef / Teichmann, Josef
en_US
ethz.identifier.url
https://openreview.net/forum?id=JFKR3WqwyXR
ethz.date.deposited
2021-11-17T08:49:40Z
ethz.source
BATCH
ethz.eth
yes
en_US
ethz.availability
Metadata only
en_US
ethz.rosetta.installDate
2021-11-18T08:34:18Z
ethz.rosetta.lastUpdated
2022-03-29T16:02:58Z
ethz.rosetta.versionExported
true
ethz.COinS
ctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.atitle=Neural%20Jump%20Ordinary%20Differential%20Equations:%20Consistent%20Continuous-Time%20Prediction%20and%20Filtering&rft.date=2021&rft.au=Herrera,%20Calypso&Krach,%20Florian&Teichmann,%20Josef&rft.genre=proceeding&rft.btitle=International%20Conference%20on%20Learning%20Representations%20(ICLR%202021)
 Search print copy at ETH Library

Files in this item

FilesSizeFormatOpen in viewer

There are no files associated with this item.

Publication type

Show simple item record