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/

Zendy Open

Presents the pdf analysis as premium feature

PDF Analysis

Insights

Presents the summarisation as premium feature

Summarisation

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

The deep supervised and reinforcement learning paradigms (among others) havethe potential to endow interactive multimodal social robots with the ability ofacquiring skills autonomously. But it is still not very clear yet how they canbe best deployed in real world applications. As a step in this direction, wepropose a deep learning-based approach for efficiently training a humanoidrobot to play multimodal games---and use the game of `Noughts & Crosses' withtwo variants as a case study. Its minimum requirements for learning to perceiveand interact are based on a few hundred example images, a few examplemultimodal dialogues and physical demonstrations of robot manipulation, andautomatic simulations. In addition, we propose novel algorithms for robustvisual game tracking and for competitive policy learning with high winningrates, which substantially outperform DQN-based baselines. While an automaticevaluation shows evidence that the proposed approach can be easily extended tonew games with competitive robot behaviours, a human evaluation with 130 humansplaying with the Pepper robot confirms that highly accurate visual perceptionis required for successful game play.

A Data-Efficient Deep Learning Approach for Deployable Multimodal Social Robots