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

The demand for more transparency of decision-making processes of deepreinforcement learning agents is greater than ever, due to their increased usein safety critical and ethically challenging domains such as autonomousdriving. In this empirical study, we address this lack of transparencyfollowing an idea that is inspired by research in the field of neuroscience. Wecharacterize the learned representations of an agent's policy network throughits activation space and perform partial network ablations to compare therepresentations of the healthy and the intentionally damaged networks. We showthat the healthy agent's behavior is characterized by a distinct correlationpattern between the network's layer activation and the performed actions duringan episode and that network ablations, which cause a strong change of thispattern, lead to the agent failing its trained control task. Furthermore, thelearned representation of the healthy agent is characterized by a distinctpattern in its activation space reflecting its different behavioral stagesduring an episode, which again, when distorted by network ablations, leads tothe agent failing its trained control task. Concludingly, we argue in favor ofa new perspective on artificial neural networks as objects of empiricalinvestigations, just as biological neural systems in neuroscientific studies,paving the way towards a new standard of scientific falsifiability with respectto research on transparency and interpretability of artificial neural networks.

How Do You Act? An Empirical Study to Understand Behavior of Deep Reinforcement Learning Agents