Training Image and Text Embedding Models
Topics: AI (Deep Learning), Document Classification, Image Search, Probably in use
The patent focusing on the advancement of search technologies through the integration of image and text analysis. This method aims to enhance search systems by training models to understand and correlate the content of images with textual search queries. This patent pertains to developing methods, systems, and apparatus, including software encoded on a computer storage medium, dedicated to improving search functionalities by jointly training an image embedding model and a text embedding model.
This patent belongs to the the topics of vector search, ranking, deep learning, scoring and image search.