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

Generating adversarial scenarios, which have the potential to fail autonomousdriving systems, provides an effective way to improve robustness. Extendingpurely data-driven generative models, recent specialized models satisfyadditional controllable requirements such as embedding a traffic sign in adriving scene by manipulating patterns implicitly in the neuron level. In thispaper, we introduce a method to incorporate domain knowledge explicitly in thegeneration process to achieve the Semantically Adversarial Generation (SAG). Tobe consistent with the composition of driving scenes, we first categorize theknowledge into two types, the property of objects and the relationship amongobjects. We then propose a tree-structured variational auto-encoder (T-VAE) tolearn hierarchical scene representation. By imposing semantic rules on theproperties of nodes and edges in the tree structure, explicit knowledgeintegration enables controllable generation. We construct a synthetic exampleto illustrate the controllability and explainability of our method in asuccinct setting. We further extend to realistic environments for autonomousvehicles: our method efficiently identifies adversarial driving scenes againstdifferent state-of-the-art 3D point cloud segmentation models and satisfies thetraffic rules specified as the explicit knowledge.

Semantically Adversarial Scenario Generation with Explicit Knowledge Guidance