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 this work, we take a representation learning perspective on hierarchicalreinforcement learning, where the problem of learning lower layers in ahierarchy is transformed into the problem of learning trajectory-levelgenerative models. We show that we can learn continuous latent representationsof trajectories, which are effective in solving temporally extended andmulti-stage problems. Our proposed model, SeCTAR, draws inspiration fromvariational autoencoders, and learns latent representations of trajectories. Akey component of this method is to learn both a latent-conditioned policy and alatent-conditioned model which are consistent with each other. Given the samelatent, the policy generates a trajectory which should match the trajectorypredicted by the model. This model provides a built-in prediction mechanism, bypredicting the outcome of closed loop policy behavior. We propose a novelalgorithm for performing hierarchical RL with this model, combining model-basedplanning in the learned latent space with an unsupervised explorationobjective. We show that our model is effective at reasoning over long horizonswith sparse rewards for several simulated tasks, outperforming standardreinforcement learning methods and prior methods for hierarchical reasoning,model-based planning, and exploration.

Self-Consistent Trajectory Autoencoder: Hierarchical Reinforcement Learning with Trajectory Embeddings