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

Training a high-dimensional simulated agent with an under-specified rewardfunction often leads the agent to learn physically infeasible strategies thatare ineffective when deployed in the real world. To mitigate these unnaturalbehaviors, reinforcement learning practitioners often utilize complex rewardfunctions that encourage physically plausible behaviors. However, a tediouslabor-intensive tuning process is often required to create hand-designedrewards which might not easily generalize across platforms and tasks. Wepropose substituting complex reward functions with "style rewards" learned froma dataset of motion capture demonstrations. A learned style reward can becombined with an arbitrary task reward to train policies that perform tasksusing naturalistic strategies. These natural strategies can also facilitatetransfer to the real world. We build upon Adversarial Motion Priors -- anapproach from the computer graphics domain that encodes a style reward from adataset of reference motions -- to demonstrate that an adversarial approach totraining policies can produce behaviors that transfer to a real quadrupedalrobot without requiring complex reward functions. We also demonstrate that aneffective style reward can be learned from a few seconds of motion capture datagathered from a German Shepherd and leads to energy-efficient locomotionstrategies with natural gait transitions.

Adversarial Motion Priors Make Good Substitutes for Complex Reward Functions