| CPC G06F 40/106 (2020.01) [G06F 3/0483 (2013.01); G06F 3/04847 (2013.01); G06F 16/483 (2019.01); G06F 16/5866 (2019.01); G06F 16/81 (2019.01); G06F 16/86 (2019.01); G06F 16/94 (2019.01); G06F 16/953 (2019.01); G06F 16/9566 (2019.01); G06F 16/9577 (2019.01); G06F 40/169 (2020.01); G06Q 30/0276 (2013.01); G06Q 30/0277 (2013.01); G06F 3/04817 (2013.01); G06F 3/0482 (2013.01); G06F 3/0489 (2013.01); G06F 16/957 (2019.01); G06F 40/20 (2020.01); G06F 40/30 (2020.01); G06N 20/00 (2019.01)] | 21 Claims |

|
1. A computer-implemented method, for similar information determination and text generation for auto-annotations of potential interest to a user within a web document, comprising:
calculating a source set of sentence embedding vectors associated with sentences coming from a plurality of source content, wherein a type of an element of the plurality of source content is selected from a group comprised of web pages, documents, and annotations of web pages and documents, using a deep learning model for sentence embeddings;
loading a web document into a web browser when a user navigates the web browser to a corresponding URL for the web document;
calculating, using the deep learning model for sentence embedding vectors, a request set of sentence embedding vectors associated with sentences coming from a plurality of request content, wherein a type of an element of the plurality of request content is selected from a group comprised of the web document, annotations of the web document, replies to the annotations of the web document, and items associated to the web document that belong to a collection that contains the web document;
retrieving through APIs a retrieved subset of the plurality of source content using the request set of sentence embedding vectors and the source set of sentence embedding vectors;
auto-annotating a subset of the plurality of request content to create auto-annotations using the retrieved subset of the plurality of source content and the plurality of request content using a deep learning model for natural language processing, wherein the subset of the plurality of request content and auto-annotations are of potential interest to the user; and
displaying the auto-annotations to the user on the web document displayed in the web browser or the annotation capable web browser.
|