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

Deep learning has become an increasingly common technique for various controlproblems, such as robotic arm manipulation, robot navigation, and autonomousvehicles. However, the downside of using deep neural networks to learn controlpolicies is their opaque nature and the difficulties of validating theirsafety. As the networks used to obtain state-of-the-art results becomeincreasingly deep and complex, the rules they have learned and how they operatebecome more challenging to understand. This presents an issue, since insafety-critical applications the safety of the control policy must be ensuredto a high confidence level. In this paper, we propose an automated black boxtesting framework based on adversarial reinforcement learning. The techniqueuses an adversarial agent, whose goal is to degrade the performance of thetarget model under test. We test the approach on an autonomous vehicle problem,by training an adversarial reinforcement learning agent, which aims to cause adeep neural network-driven autonomous vehicle to collide. Two neural networkstrained for autonomous driving are compared, and the results from the testingare used to compare the robustness of their learned control policies. We showthat the proposed framework is able to find weaknesses in both control policiesthat were not evident during online testing and therefore, demonstrate asignificant benefit over manual testing methods.

Training Adversarial Agents to Exploit Weaknesses in Deep Control Policies