Domain knowledge learning techniques for natural language generation

A system, comprising:

• at least one computer hardware processor;
• and
• at least one non-transitory computer-readable storage medium storing processor executable instructions that, when executed by the at least one computer hardware processor, cause the at least one computer hardware processor to perform:
• a method for identifying vocabulary associated with a plurality of semantic objects used for generating natural language text with a natural language generation (NLG) system, the plurality of semantic objects including a first semantic object having a first set of ordered attributes and second semantic object having a second set of ordered attributes, the method comprising:
• obtaining at least one text corpus comprising a plurality of text segments;
• identifying, from among the plurality of text segments and using at least one first machine learning classifier, groups of text segments corresponding to respective semantic objects in the plurality of semantic objects, the groups of text segments including a first group of text segments corresponding to the first semantic object and a second group of text segments corresponding to the second semantic object;
• identifying, from the groups of text segments and using at least one second machine learning classifier, a plurality of vocabularies for the plurality of semantic objects, the identifying comprising:
• identifying, using the first group of text segments and the at least one second machine learning classifier, a first vocabulary for the first semantic object, the first vocabulary containing multiple alternative words for each of at least some attributes in the first set of ordered attributes;
• and
• identifying, using the second group of text segments and the at least one second machine learning classifier, a second vocabulary for the second semantic object, the second vocabulary containing multiple alternative words for each of at least some attributes in the second set of ordered attributes;
• generating natural language text using the NLG system, the plurality of vocabularies, and the plurality of semantic objects, the generated natural language text comprising a first text segment generated using the first semantic object and containing at least some words in the first vocabulary and a second text segment generated using the second semantic object and containing at least some words in the second vocabulary;
• and
• outputting the generated natural language text.

At least one non-transitory computer-readable storage medium storing processor executable instructions that, when executed by at least one computer hardware processor, cause the at least one computer hardware processor to perform:

• a method for identifying vocabulary associated with a plurality of semantic objects used for generating natural language text with a natural language generation (NLG) system, the plurality of semantic objects including a first semantic object having a first set of ordered attributes and second semantic object having a second set of ordered attributes, the method comprising:
• obtaining at least one text corpus comprising a plurality of text segments;
• identifying, from among the plurality of text segments and using at least one first machine learning classifier, groups of text segments corresponding to respective semantic objects in the plurality of semantic objects, the groups of text segments including a first group of text segments corresponding to the first semantic object and a second group of text segments corresponding to the second semantic object;
• identifying, from the groups of text segments and using at least one second machine learning classifier, a plurality of vocabularies for the plurality of semantic objects, the identifying comprising:
• identifying, using the first group of text segments and the at least one second machine learning classifier, a first vocabulary for the first semantic object, the first vocabulary containing multiple alternative words for each of at least some attributes in the first set of ordered attributes;
• and
• identifying, using the second group of text segments and the at least one second machine learning classifier, a second vocabulary for the second semantic object, the second vocabulary containing multiple alternative words for each of at least some attributes in the second set of ordered attributes;
• generating natural language text using the NLG engine, the plurality of vocabularies, and the plurality of semantic objects, the generated natural language text comprising a first text segment generated using the first semantic object and containing at least some words in the first vocabulary and a second text segment generated using the second semantic object and containing at least some words in the second vocabulary;
• and
• outputting the generated natural language text.