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

In robotics, catastrophic interference continues to restrain policy trainingacross environments. Efforts to combat catastrophic interference to date focuson novel neural architectures or training methods, with a recent emphasis onpolicies with good initial settings that facilitate training in newenvironments. However, none of these methods to date have taken into accounthow the physical architecture of the robot can obstruct or facilitatecatastrophic interference, just as the choice of neural architecture can. Inprevious work we have shown how aspects of a robot's physical structure(specifically, sensor placement) can facilitate policy learning by increasingthe fraction of optimal policies for a given physical structure. Here we showfor the first time that this proxy measure of catastrophic interferencecorrelates with sample efficiency across several search methods, proving thatfavorable loss landscapes can be induced by the correct choice of physicalstructure. We show that such structures can be found via co-optimization --optimization of a robot's structure and control policy simultaneously --yielding catastrophic interference resistant robot structures and policies, andthat this is more efficient than control policy optimization alone. Finally, weshow that such structures exhibit sensor homeostasis across environments andintroduce this as the mechanism by which certain robots overcome catastrophicinterference.

A good body is all you need: avoiding catastrophic interference via agent architecture search