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

Pre-training the embedding of a location generated from human mobility datahas become a popular method for location based services. In practice, modelingthe location embedding is too expensive, due to the large number of locationsto be trained in situations with fine-grained resolution or extensive targetregions. Previous studies have handled less than ten thousand distinctlocations, which is insufficient in the real-world applications. To tackle thisproblem, we propose a Geo-Tokenizer, designed to efficiently reduce the numberof locations to be trained by representing a location as a combination ofseveral grids at different scales. In the Geo-Tokenizer, a grid at a largerscale shares the common set of grids at smaller scales, which is a key factorin reducing the size of the location vocabulary. The sequences of locationspreprocessed with the Geo-Tokenizer are utilized by a causal location embeddingmodel to capture the temporal dependencies of locations. This model dynamicallycalculates the embedding vector of a target location, which varies depending onits trajectory. In addition, to efficiently pre-train the location embeddingmodel, we propose the Hierarchical Auto-regressive Location Model objective toeffectively train decomposed locations in the Geo-Tokenizer. We conductedexperiments on two real-world user trajectory datasets using our pre-trainedlocation model. The experimental results show that our model significantlyimproves the performance of downstream tasks with fewer model parameterscompared to existing location embedding methods.

Pre-training Contextual Location Embeddings in Personal Trajectories via Efficient Hierarchical Location Representations