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

Humans are masters at quickly learning many complex tasks, relying on anapproximate understanding of the dynamics of their environments. In much thesame way, we would like our learning agents to quickly adapt to new tasks. Inthis paper, we explore how model-based Reinforcement Learning (RL) canfacilitate transfer to new tasks. We develop an algorithm that learns anaction-conditional, predictive model of expected future observations, rewardsand values from which a policy can be derived by following the gradient of theestimated value along imagined trajectories. We show how robust policyoptimization can be achieved in robot manipulation tasks even with approximatemodels that are learned directly from vision and proprioception. We evaluatethe efficacy of our approach in a transfer learning scenario, re-usingpreviously learned models on tasks with different reward structures and visualdistractors, and show a significant improvement in learning speed compared tostrong off-policy baselines. Videos with results can be found athttps://sites.google.com/view/ivg-corl19

Imagined Value Gradients: Model-Based Policy Optimization with Transferable Latent Dynamics Models