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

Validating the safety of autonomous systems generally requires the use ofhigh-fidelity simulators that adequately capture the variability of real-worldscenarios. However, it is generally not feasible to exhaustively search thespace of simulation scenarios for failures. Adaptive stress testing (AST) is amethod that uses reinforcement learning to find the most likely failure of asystem. AST with a deep reinforcement learning solver has been shown to beeffective in finding failures across a range of different systems. Thisapproach generally involves running many simulations, which can be veryexpensive when using a high-fidelity simulator. To improve efficiency, wepresent a method that first finds failures in a low-fidelity simulator. It thenuses the backward algorithm, which trains a deep neural network policy using asingle expert demonstration, to adapt the low-fidelity failures tohigh-fidelity. We have created a series of autonomous vehicle validation casestudies that represent some of the ways low-fidelity and high-fidelitysimulators can differ, such as time discretization. We demonstrate in a varietyof case studies that this new AST approach is able to find failures withsignificantly fewer high-fidelity simulation steps than are needed when justrunning AST directly in high-fidelity. As a proof of concept, we alsodemonstrate AST on NVIDIA's DriveSim simulator, an industry state-of-the-arthigh-fidelity simulator for finding failures in autonomous vehicles.

Finding Failures in High-Fidelity Simulation using Adaptive Stress Testing and the Backward Algorithm