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 options framework in Hierarchical Reinforcement Learning breaks downoverall goals into a combination of options or simpler tasks and associatedpolicies, allowing for abstraction in the action space. Ideally, these optionscan be reused across different higher-level goals; indeed, such reuse isnecessary to realize the vision of a continual learning agent that caneffectively leverage its prior experience. Previous approaches have onlyproposed limited forms of transfer of prelearned options to new task settings.We propose a novel option indexing approach to hierarchical learning (OI-HRL),where we learn an affinity function between options and the items present inthe environment. This allows us to effectively reuse a large library ofpretrained options, in zero-shot generalization at test time, by restrictinggoal-directed learning to only those options relevant to the task at hand. Wedevelop a meta-training loop that learns the representations of options andenvironments over a series of HRL problems, by incorporating feedback about therelevance of retrieved options to the higher-level goal. We evaluate OI-HRL intwo simulated settings - the CraftWorld and AI2THOR environments - and showthat we achieve performance competitive with oracular baselines, andsubstantial gains over a baseline that has the entire option pool available forlearning the hierarchical policy.

Matching options to tasks using Option-Indexed Hierarchical Reinforcement Learning