Research on Visual Similarity in Adult Word Extension and Artificial Word Learning

ResearchPublished: Source: Hugging Face Papers

#Hugging Face

Key point

Reanalysis using DINOv2 showed that visual distance between objects is associated with adult word extension.

Research on visual similarity in adult word extension and artificial word learning was featured on Hugging Face Daily Papers. The study reanalyzed trial-level data from a large adult word-extension study (Lewis & Frank, 2018).

Using the space of a frozen self-supervised visual encoder, DINOv2, the researchers computed the distance of each test image to the shown examples. In a study of 520 people, distance was associated with extension (log odds -3.23 per standard deviation). An exploratory post hoc analysis suggested that this association persisted within the taught domain after controlling for taxonomic level and teaching condition.

Example composition also mattered; after three examples, adults extended a word to 1. The paper received 2 upvotes on Hugging Face Daily Papers.

This item is a summary. The full, checkable version is at the original source:

Read the original at Hugging Face Papers ↗