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 Plus

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

Zendy Tools

Zendy Open

Humans should be able work more effectively with artificialintelligence-based systems when they can predict likely failures and formuseful mental models of how the systems work. We conducted a study of human'smental models of artificial intelligence systems using a high-performing imageclassifier, focusing on participants' ability to predict the classificationresult for a particular image. Participants viewed individual labeled images inone of two classes and then tried to predict whether the classifier would labelthem correctly. In this experiment we explored the effect of givingparticipants additional information about an image's nearest neighbors in aspace representing the otherwise uninterpretable features extracted by thelower layers of the classifier's neural network. We found that providing thisinformation did increase participants' prediction performance, and that theperformance improvement could be related to the neighbor images' similarity tothe target image. We also found indications that the presentation of thisinformation may influence people's own classification of the target image --that is, rather than just anthropomorphizing the system, in some cases thehumans become "mechanomorphized" in their judgements.

Won't you see my neighbor?: User predictions, mental models, and similarity-based explanations of AI classifiers