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

Animals need to devise strategies to maximize returns while interacting withtheir environment based on incoming noisy sensory observations. Task-relevantstates, such as the agent's location within an environment or the presence of apredator, are often not directly observable but must be inferred usingavailable sensory information. Successor representations (SR) have beenproposed as a middle-ground between model-based and model-free reinforcementlearning strategies, allowing for fast value computation and rapid adaptationto changes in the reward function or goal locations. Indeed, recent studiessuggest that features of neural responses are consistent with the SR framework.However, it is not clear how such representations might be learned and computedin partially observed, noisy environments. Here, we introduce a neurallyplausible model using distributional successor features, which builds on thedistributed distributional code for the representation and computation ofuncertainty, and which allows for efficient value function computation inpartially observed environments via the successor representation. We show thatdistributional successor features can support reinforcement learning in noisyenvironments in which direct learning of successful policies is infeasible.

A neurally plausible model learns successor representations in partially observable environments