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

Demystifying the interactions among multiple agents from their pasttrajectories is fundamental to precise and interpretable trajectory prediction.However, previous works only consider pair-wise interactions with limitedrelational reasoning. To promote more comprehensive interaction modeling forrelational reasoning, we propose GroupNet, a multiscale hypergraph neuralnetwork, which is novel in terms of both interaction capturing andrepresentation learning. From the aspect of interaction capturing, we propose atrainable multiscale hypergraph to capture both pair-wise and group-wiseinteractions at multiple group sizes. From the aspect of interactionrepresentation learning, we propose a three-element format that can be learntend-to-end and explicitly reason some relational factors including theinteraction strength and category. We apply GroupNet into both CVAE-basedprediction system and previous state-of-the-art prediction systems forpredicting socially plausible trajectories with relational reasoning. Tovalidate the ability of relational reasoning, we experiment with syntheticphysics simulations to reflect the ability to capture group behaviors, reasoninteraction strength and interaction category. To validate the effectiveness ofprediction, we conduct extensive experiments on three real-world trajectoryprediction datasets, including NBA, SDD and ETH-UCY; and we show that withGroupNet, the CVAE-based prediction system outperforms state-of-the-artmethods. We also show that adding GroupNet will further improve the performanceof previous state-of-the-art prediction systems.

GroupNet: Multiscale Hypergraph Neural Networks for Trajectory Prediction with Relational Reasoning