Generation of text classifier training data

A system for generating a text classifier for a document corpus, the system comprising:

• one or more processors;
• and
• one or more memory devices coupled to the one or more processors, the one or more memory devices storing instructions that are executable by the one or more processors to perform operations comprising:
• receiving input designating a term of interest in a document of the document corpus;
• determining a target term embedding representing the term of interest;
• determining a set of candidate word groups of the document corpus, wherein the set of candidate word groups corresponds to a subset of word groups of the document corpus that include the term of interest or a term that is semantically similar to the term of interest based on distances between the target term embedding and term embeddings representing terms from the document corpus;
• determining a target context embedding representing a target word group, the target word group including the term of interest and context words located in the document proximate to the term of interest;
• identifying, from among the document corpus, a first candidate word group that is semantically similar to the target word group based on a distance between the target context embedding and a context embedding of the first candidate word group;
• identifying, from among the set of candidate word groups, a second candidate word group that is semantically dissimilar to the target word group based on a distance between the target context embedding and a context embedding of the second candidate word group;
• presenting, via a user interface, output including the first candidate word group and the second candidate word group;
• receiving user input via the user interface, the user input identifying at least a portion of the first candidate word group as associated with a first label and identifying at least a portion of the second candidate word group as not associated with the first label;
• and
• generating labeled training data based on the user input to train the text classifier.

A computer-readable storage device storing instructions that are executable by a processor to perform operations including:

• receiving input designating a term of interest in a document of a document corpus;
• determining a target term embedding representing the term of interest;
• determining a set of candidate word groups of the document corpus, wherein the set of candidate word groups corresponds to a subset of word groups of the document corpus that include the term of interest or a term that is semantically similar to the term of interest based on distances between the target term embedding and term embeddings representing terms from the document corpus;
• determining a target context embedding representing a target word group, the target word group including the term of interest and context words located in the document proximate to the term of interest;
• identifying, from among the document corpus, a first candidate word group that is semantically similar to the target word group based on a distance between the target context embedding and a context embedding of the first candidate word group;
• identifying, from among the set of candidate word groups, a second candidate word group that is semantically dissimilar to the target word group based on a distance between the target context embedding and a context embedding of the second candidate word group;
• presenting, via a user interface, output including the first candidate word group and the second candidate word group;
• receiving user input via the user interface, the user input identifying at least a portion of the first candidate word group as associated with a first label and identifying at least a portion of the second candidate word group as not associated with the first label;
• and
• generating labeled training data based on the user input to train a text classifier.