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

Systems consisting of interacting agents are prevalent in the world, rangingfrom dynamical systems in physics to complex biological networks. To buildsystems which can interact robustly in the real world, it is thus important tobe able to infer the precise interactions governing such systems. Existingapproaches typically discover such interactions by explicitly modeling thefeed-forward dynamics of the trajectories. In this work, we propose NeuralInteraction Inference with Potentials (NIIP) as an alternative approach todiscover such interactions that enables greater flexibility in trajectorymodeling: it discovers a set of relational potentials, represented as energyfunctions, which when minimized reconstruct the original trajectory. NIIPassigns low energy to the subset of trajectories which respect the relationalconstraints observed. We illustrate that with these representations NIIPdisplays unique capabilities in test-time. First, it allows trajectorymanipulation, such as interchanging interaction types across separately trainedmodels, as well as trajectory forecasting. Additionally, it allows addingexternal hand-crafted potentials at test-time. Finally, NIIP enables thedetection of out-of-distribution samples and anomalies without explicittraining. Website: https://energy-based-model.github.io/interaction-potentials.

Inferring Relational Potentials in Interacting Systems