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

Sensors are commonly deployed to perceive the environment. However, due tothe high cost, sensors are usually sparsely deployed. Kriging is the tailoredtask to infer the unobserved nodes (without sensors) using the observed sourcenodes (with sensors). The essence of kriging task is transferability. Recently,several inductive spatio-temporal kriging methods have been proposed based ongraph neural networks, being trained based on a graph built on top of observednodes via pretext tasks such as masking nodes out and reconstructing them.However, the graph in training is inevitably much sparser than the graph ininference that includes all the observed and unobserved nodes. The learnedpattern cannot be well generalized for inference, denoted as graph gap. Toaddress this issue, we first present a novel Increment training strategy:instead of masking nodes (and reconstructing them), we add virtual nodes intothe training graph so as to mitigate the graph gap issue naturally.Nevertheless, the empty-shell virtual nodes without labels could havebad-learned features and lack supervision signals. To solve these issues, wepair each virtual node with its most similar observed node and fuse theirfeatures together; to enhance the supervision signal, we construct reliablepseudo labels for virtual nodes. As a result, the learned pattern of virtualnodes could be safely transferred to real unobserved nodes for reliablekriging. We name our new Kriging model with Increment Training Strategy asKITS. Extensive experiments demonstrate that KITS consistently outperformsexisting kriging methods by large margins, e.g., the improvement over MAE scorecould be as high as 18.33%.

KITS: Inductive Spatio-Temporal Kriging with Increment Training Strategy