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

Evaluating the performance of autonomous vehicles (AV) and their complexsubsystems to high precision under naturalistic circumstances remains achallenge, especially when failure or dangerous cases are rare. Rarity does notonly require an enormous sample size for a naive method to achieve highconfidence estimation, but it also causes dangerous underestimation of the truefailure rate and it is extremely hard to detect. Meanwhile, thestate-of-the-art approach that comes with a correctness guarantee can onlycompute an upper bound for the failure rate under certain conditions, whichcould limit its practical uses. In this work, we present Deep ImportanceSampling (Deep IS) framework that utilizes a deep neural network to obtain anefficient IS that is on par with the state-of-the-art, capable of reducing therequired sample size 43 times smaller than the naive sampling method to achieve10% relative error and while producing an estimate that is much lessconservative. Our high-dimensional experiment estimating the misclassificationrate of one of the state-of-the-art traffic sign classifiers further revealsthat this efficiency still holds true even when the target is very small,achieving over 600 times efficiency boost. This highlights the potential ofDeep IS in providing a precise estimate even against high-dimensionaluncertainties.

Test Against High-Dimensional Uncertainties: Accelerated Evaluation of Autonomous Vehicles with Deep Importance Sampling