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

Navigation is crucial for animal behavior and is assumed to require aninternal representation of the external environment, termed a cognitive map.The precise form of this representation is often considered to be a metricrepresentation of space. An internal representation, however, is judged by itscontribution to performance on a given task, and may thus vary betweendifferent types of navigation tasks. Here we train a recurrent neural networkthat controls an agent performing several navigation tasks in a simpleenvironment. To focus on internal representations, we split learning into atask-agnostic pre-training stage that modifies internal connectivity and atask-specific Q learning stage that controls the network's output. We show thatpre-training shapes the attractor landscape of the networks, leading to eithera continuous attractor, discrete attractors or a disordered state. Thesestructures induce bias onto the Q-Learning phase, leading to a performancepattern across the tasks corresponding to metric and topological regularities.By combining two types of networks in a modular structure, we could get betterperformance for both regularities. Our results show that, in recurrentnetworks, inductive bias takes the form of attractor landscapes -- which can beshaped by pre-training and analyzed using dynamical systems methods.Furthermore, we demonstrate that non-metric representations are useful fornavigation tasks, and their combination with metric representation leads toflexibile multiple-task learning.

Implementing Inductive bias for different navigation tasks through diverse RNN attractors