Abstract: Aspects of the invention may comprise a method for improving execution speed of a semantic query on a database by executing the semantic query on a subset of relevant documents in the database. The method may provide for improving execution speed without sacrificing accuracy because the subset of documents comprises documents relevant to the semantic query. Aspects of the invention may further provide for identifying key concepts in a semantic query and generating a filter query based on the identified key concepts. The filter query may be used to select the subset of relevant documents. Aspects further provide for generating one or more keyword queries based on a key concept and at least one search operator. The filter query may incorporate at least one keyword query. Aspects of the invention further comprise implementing entity recognition, metadata filtering, and essentialness to improve the selected subset.

Abstract: Document difference viewing and navigation are disclosed that enables automatic comparison and easy viewing of the difference between many documents (e.g., four or more, dozens, hundreds, or thousands of documents) for, for example, document review processes. The system obtains and reads a set of documents containing text. The system identifies shared text and differences by location within the shared text among different documents in the set of documents. It sorts the documents in the set of documents into groups of different text variation versions and generates a shared difference document. The shared difference document is displayed in a first panel of a user interface and displays shared text as static text and areas of difference in highlighted regions. When a user selects of a given area of difference, the text variations and a list of documents containing one or more corresponding text variations are displayed in a second panel.

Abstract: Document difference viewing and navigation are disclosed that enables automatic comparison and easy viewing of the difference between many documents (e.g., four or more, dozens, hundreds, or thousands of documents) for, for example, document review processes. The system receives and reads a set of documents containing text. The system identifies shared text and differences by location within the shared text among different documents in the set of documents. It sorts the documents in the set of documents into groups of different text variation versions and generates a shared difference document. The shared difference document is displayed in a first panel of a user interface and displays shared text as static text and areas of difference in highlighted regions. When a user selects of a given area of difference, the text variations and a list of documents containing one or more corresponding text variations are displayed in a second panel.

Abstract: Distributed email threading is disclosed. Emails are grouped into smaller related groups, as opposed to one large group, so that the groups are processed serially or in parallel to save processing power and time. Each email message group is handled in an independent job, which is distributable among any number of servers. A bipartite graph of email messages and message identifiers (such as subject line, recipient address, etc.) is created and used to determine connected sub-groups within the graph and, ultimately, which messages are connected to each other as part of an email thread. The sub-groups are processed independently for efficiency.

Abstract: Distributed email threading is disclosed. Emails are grouped into smaller related groups, as opposed to one large group, so that the groups are processed serially or in parallel to save processing power and time. Each email message group is handled in an independent job, which is distributable among any number of servers. A bipartite graph of email messages and message identifiers (such as subject line, recipient address, etc.) is created and used to determine connected sub-groups within the graph and, ultimately, which messages are connected to each other as part of an email thread. The sub-groups are processed independently for efficiency.

Abstract: Document difference viewing and navigation are disclosed that enables automatic comparison and easy viewing of the difference between many documents (e.g., four or more, dozens, hundreds, or thousands of documents) for, for example, document review processes. The system receives and reads a set of documents containing text. The system identifies shared text and differences by location within the shared text among different documents in the set of documents. It sorts the documents in the set of documents into groups of different text variation versions and generates a shared difference document. The shared difference document is displayed in a first panel of a user interface and displays shared text as static text and areas of difference in highlighted regions. When a user selects of a given area of difference, the text variations and a list of documents containing one or more corresponding text variations are displayed in a second panel.