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/

Zendy Plus

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

Zendy Tools

Zendy Open

Some group activities, such as team sports and choreographed dances, involveclosely coupled interaction between participants. Here we investigate the tasksof inferring and predicting participant behavior, in terms of motion paths andactions, under such conditions. We narrow the problem to that of estimating howa set target participants react to the behavior of other observed participants.Our key idea is to model the spatio-temporal relations among participants in amanner that is robust to error accumulation during frame-wise inference andprediction. We propose a novel Entry-Flipped Transformer (EF-Transformer),which models the relations of participants by attention mechanisms on bothspatial and temporal domains. Unlike typical transformers, we tackle theproblem of error accumulation by flipping the order of query, key, and valueentries, to increase the importance and fidelity of observed features in thecurrent frame. Comparative experiments show that our EF-Transformer achievesthe best performance on a newly-collected tennis doubles dataset, a Ceilidhdance dataset, and two pedestrian datasets. Furthermore, it is alsodemonstrated that our EF-Transformer is better at limiting accumulated errorsand recovering from wrong estimations.

Entry-Flipped Transformer for Inference and Prediction of Participant Behavior