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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The problem of balancing conflicting needs is fundamental to intelligence.Standard reinforcement learning algorithms maximize a scalar reward, whichrequires combining different objective-specific rewards into a single number.Alternatively, different objectives could also be combined at the level ofaction value, such that specialist modules responsible for different objectivessubmit different action suggestions to a decision process, each based onrewards that are independent of one another. In this work, we explore thepotential benefits of this alternative strategy. We investigate a biologicallyrelevant multi-objective problem, the continual homeostasis of a set ofvariables, and compare a monolithic deep Q-network to a modular network with adedicated Q-learner for each variable. We find that the modular agent: a)requires minimal exogenously determined exploration; b) has improved sampleefficiency; and c) is more robust to out-of-domain perturbation.

Modularity benefits reinforcement learning agents with competing homeostatic drives