Converting from compressed language to natural language

A system comprising:

• at least one processor;
• a set controller configured to generate a token set;
• a name controller configured to generate a name string;
• an element language model configured to identify compressed language word sequences;
• a corpus language model configured to identify natural language word sequences;
• and
• a server application executing on the at least one processor and configured for:
• receiving an element string, wherein the element string is in a compressed language format and is extracted from a document in a structured document language,
• tokenizing the element string to form a plurality of element tokens,
• determining that an element token of the plurality of element tokens does not match a cognate token, of a corpus language model for a natural language, as a word of the natural language,
• in response to determining that the element token does not match the cognate token:
• generating a plurality of matched tokens from the element token using an element language model for a compressed language, wherein the compressed language is a compressed language version of the natural language with one or more characters removed from one or more words of the natural language,
• reranking the plurality of matched tokens to form a plurality of reranked tokens using the corpus language model,
• generating, by the set controller, the token set from the plurality of element tokens using the element language model, the corpus language model, and the plurality of reranked tokens,
• generating, by the name controller, the name string from a plurality of token sets comprising the token set, and, wherein the name string is in a natural language format,
• presenting the name string in response to a user input after being automatically generated.

A method comprising:

• receiving a structured document, the structured document comprising an element string,
• wherein the element string is in a compressed language format,
• wherein the element string is tokenized to form a plurality of element tokens,
• wherein an element token of the plurality of element tokens does not match a cognate token, of a corpus language model for a natural language, as a word of the natural language,
• wherein, in response to determining that the element token does not match the cognate token:
• a plurality of matched tokens is generated from the element token using an element language model for a compressed language, wherein the compressed language is a compressed language version of the natural language with one or more characters removed from one or more words of the natural language, and
• the plurality of matched tokens is reranked to form a plurality of reranked tokens using the corpus language model,
• wherein a set controller is applied to the plurality of reranked tokens to generate a token set, and
• wherein a name controller is applied to a plurality of token sets, comprising the token set, to generate a name string, from the element string, in a natural language format;
• and
• displaying the structured document with the name string in the natural language format in response to a user input.