[ABE-L] Fwd: Colóquio do IME - 10/12, 14h - Ying Sun

[Márcia Branco]

Divulgando Colóquio do IME-USP com a conferência da Profa Ying Sun -
 https://cemse.kaust.edu.sa/profiles/ying-sun


---------- Forwarded message ---------
De: Colóquio do IME <coloquio em ime.usp.br>
Date: seg., 2 de dez. de 2024, 08:00
Subject: Colóquio do IME - 10/12, 14h -


Bom dia,

É com grande satisfação que convidamos a comunidade para a última palestra
deste ano no Colóquio do IME, que ocorre *excepcionalmente na terça* da
semana que vem.

Ying Sun*, *King Abdullah University of Science and Technology
*10 de dezembro, às 14h, no Auditório Antonio Gilioli*
*Título:* Spatio-temporal DeepKriging for Interpolation and Probabilistic
Forecasting
*Resumo:* Gaussian processes (GP) and Kriging are widely used in
traditional spatio-temporal modeling and prediction. These techniques
typically presuppose that the data are observed from a stationary GP with a
parametric covariance structure. However, processes in real-world
applications often exhibit non-Gaussianity and nonstationarity. Moreover,
likelihood-based inference for GPs is computationally expensive and thus
prohibitive for large datasets. In this paper, we propose a deep neural
network (DNN) based two-stage model for spatio-temporal interpolation and
forecasting. Interpolation is performed in the first step, which utilizes a
dependent DNN with the embedding layer constructed with spatio-temporal
basis functions. For the second stage, we use Long Short-Term Memory (LSTM)
and convolutional LSTM to forecast future observations at a given location.
We adopt the quantile-based loss function in the DNN to provide
probabilistic forecasting. Compared to Kriging, the proposed method does
not require specifying covariance functions or making stationarity
assumptions and is computationally efficient. Therefore, it is suitable for
large-scale prediction of complex spatio-temporal processes. We apply our
method to monthly PM2.5 data at more than 200,000 space–time locations for
fast imputation of missing values and forecasts with uncertainties.

Participem e ajudem a divulgar! Mais informações sobre o Colóquio do IME em
https://www.ime.usp.br/coloquio/.

Saudações,
Equipe do Colóquio do IME

-- 
Para cancelar a inscrição neste grupo, envie um e-mail para
g-mae+unsubscribe em ime.usp.br
-------------- Próxima Parte ----------
Um anexo em HTML foi limpo...
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20241203/b0dc60f6/attachment.htm>