Apr 18, 2023 · We propose MERU, a contrastive model that yields hyperbolic representations of images and text. Hyperbolic spaces have suitable geometric properties to embed ...
Visual and linguistic concepts naturally organize themselves in a hierarchy, where a textual con- cept “dog” entails all images that contain dogs.
In this paper, we learn large-scale image-text representations (MERU) to capture the visual-semantic hierarchy underlying images and text. Our key innovation is ...
Jul 23, 2023 · We propose MERU, a contrastive model that yields hyperbolic representations of images and text. Hyperbolic spaces have suitable geometric ...
Aug 19, 2024 · We propose \textbf{HYDEN}, a novel hyperbolic density embedding based image-text representation learning approach tailored for specific medical domain data.
Jan 18, 2024 · Visual and linguistic concepts naturally organize themselves in a hierarchy, where a textual concept "dog" entails all images that contain ...
We observe that the degree of hyperbolicity in image datasets is quite high, as the obtained δrel are significantly closer to 0 than to 1. (which would indicate ...