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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Capturing multimodal natures is essential for stochastic pedestriantrajectory prediction, to infer a finite set of future trajectories. Theinferred trajectories are based on observation paths and the latent vectors ofpotential decisions of pedestrians in the inference step. However, stochasticapproaches provide varying results for the same data and parameter settings,due to the random sampling of the latent vector. In this paper, we analyze theproblem by reconstructing and comparing probabilistic distributions fromprediction samples and socially-acceptable paths, respectively. Through thisanalysis, we observe that the inferences of all stochastic models are biasedtoward the random sampling, and fail to generate a set of realistic paths fromfinite samples. The problem cannot be resolved unless an infinite number ofsamples is available, which is infeasible in practice. We introduce that theQuasi-Monte Carlo (QMC) method, ensuring uniform coverage on the samplingspace, as an alternative to the conventional random sampling. With the samefinite number of samples, the QMC improves all the multimodal predictionresults. We take an additional step ahead by incorporating a learnable samplingnetwork into the existing networks for trajectory prediction. For this purpose,we propose the Non-Probability Sampling Network (NPSN), a very small network(~5K parameters) that generates purposive sample sequences using the past pathsof pedestrians and their social interactions. Extensive experiments confirmthat NPSN can significantly improve both the prediction accuracy (up to 60%)and reliability of the public pedestrian trajectory prediction benchmark. Codeis publicly available at https://github.com/inhwanbae/NPSN .

Non-Probability Sampling Network for Stochastic Human Trajectory Prediction