From what I glean, it seems that you are generating patent embeddings directly. Is this correct? and if so, are you then forced to rerun the model every time new patents are published?
Is it possible to generate new embeddings using unpublished patent data like an abstract, etc?
From what I glean, it seems that you are generating patent embeddings directly. Is this correct? and if so, are you then forced to rerun the model every time new patents are published?
Is it possible to generate new embeddings using unpublished patent data like an abstract, etc?