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

This work addresses the problem of sensing the world: how to learn amultimodal representation of a reinforcement learning agent's environment thatallows the execution of tasks under incomplete perceptual conditions. Toaddress such problem, we argue for hierarchy in the design of representationmodels and contribute with a novel multimodal representation model, MUSE. Theproposed model learns hierarchical representations: low-level modality-specificrepresentations, encoded from raw observation data, and a high-level multimodalrepresentation, encoding joint-modality information to allow robust stateestimation. We employ MUSE as the sensory representation model of deepreinforcement learning agents provided with multimodal observations in Atarigames. We perform a comparative study over different designs of reinforcementlearning agents, showing that MUSE allows agents to perform tasks underincomplete perceptual experience with minimal performance loss. Finally, weevaluate the performance of MUSE in literature-standard multimodal scenarioswith higher number and more complex modalities, showing that it outperformsstate-of-the-art multimodal variational autoencoders in single andcross-modality generation.

How to Sense the World: Leveraging Hierarchy in Multimodal Perception for Robust Reinforcement Learning Agents