Abstract: An embedding associated with a query comprising one or more query terms is determined to be not stored in memory. An embedding is generated for the one or more query terms. One or more web documents that are similar to the generated embedding are determined. One or more content cards associated with the one or more web documents that are determined to be similar to the generated embedding are provided in a content feed.

Abstract: A curation score associated with an entity is determined. A good interest probability value associated with the entity is determined. A content feed for a user that includes one or more web documents is generated based in part on the curation score and the good interest probability value.

Abstract: Techniques for enhanced search for generating a content feed are disclosed. In some embodiments, a system/process/computer program product for enhanced search for generating a content feed includes determining a plurality of interests for a user, wherein the user is associated with a user account; searching one or more websites based on the plurality of interests associated with the user; generating an index that includes a plurality of web documents and relationships between each of the plurality of web documents; and generating a content feed that includes at least a subset of the plurality of web documents based on a ranking, wherein the ranking is based on the plurality of interests associated with the user.

Abstract: Techniques for providing an enhanced search to generate a feed based on a user's interests are disclosed. In some embodiments, a system/process/computer program product for providing an enhanced search to generate a feed based on a user's interests includes receiving a plurality of interests associated with a user, searching online content including one or more websites (e.g., news or other content websites, social networking sites, and/or other online content) based on the plurality of interests associated with the user, receiving a plurality of web documents (e.g., links to websites, social networking sites, and other online content) based on the search for online content, ranking the plurality of web documents based on a document score and a user signal, and generating a content feed that includes at least a subset of the plurality of web documents based on the ranking.

Abstract: A computer-implemented method for classifying documents for data loss prevention may include 1) identifying training documents for a machine learning classifier configured for data loss prevention, 2) performing a semantic analysis on training documents to identify topics within the set training documents, 3) applying a similarity metric to the topics to identify at least one unrelated topic with a similarity to the other topics within the plurality of topics, as determined by the similarity metric, that falls below a similarity threshold, 4) identifying, based on the semantic analysis, at least one irrelevant training document within the set of training documents in which a predominance of the unrelated topic is above a predominance threshold, and 5) excluding the irrelevant training document from the set of training documents based on the predominance of the unrelated topic within the irrelevant training document. Various other methods, systems, and computer-readable media are also disclosed.