English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Dealing with missing values and incomplete time series is a labor-intensive,tedious, inevitable task when handling data coming from real-worldapplications. Effective spatio-temporal representations would allow imputationmethods to reconstruct missing temporal data by exploiting information comingfrom sensors at different locations. However, standard methods fall short incapturing the nonlinear time and space dependencies existing within networks ofinterconnected sensors and do not take full advantage of the available - andoften strong - relational information. Notably, most state-of-the-artimputation methods based on deep learning do not explicitly model relationalaspects and, in any case, do not exploit processing frameworks able toadequately represent structured spatio-temporal data. Conversely, graph neuralnetworks have recently surged in popularity as both expressive and scalabletools for processing sequential data with relational inductive biases. In thiswork, we present the first assessment of graph neural networks in the contextof multivariate time series imputation. In particular, we introduce a novelgraph neural network architecture, named GRIN, which aims at reconstructingmissing data in the different channels of a multivariate time series bylearning spatio-temporal representations through message passing. Empiricalresults show that our model outperforms state-of-the-art methods in theimputation task on relevant real-world benchmarks with mean absolute errorimprovements often higher than 20%.

Filling the G_ap_s: Multivariate Time Series Imputation by Graph Neural Networks