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

A successful robot-assisted feeding system requires bite acquisition of awide variety of food items. It must adapt to changing user food preferencesunder uncertain visual and physical environments. Different food items indifferent environmental conditions require different manipulation strategiesfor successful bite acquisition. Therefore, a key challenge is how to handlepreviously unseen food items with very different success rate distributionsover strategy. Combining low-level controllers and planners into discreteaction trajectories, we show that the problem can be represented using a linearcontextual bandit setting. We construct a simulated environment using a doublyrobust loss estimate from previously seen food items, which we use to tune theparameters of off-the-shelf contextual bandit algorithms. Finally, wedemonstrate empirically on a robot-assisted feeding system that, even startingwith a model trained on thousands of skewering attempts on dissimilarpreviously seen food items, $\epsilon$-greedy and LinUCB algorithms can quicklyconverge to the most successful manipulation strategy.

Adaptive Robot-Assisted Feeding: An Online Learning Framework for Acquiring Previously Unseen Food Items