Multi-lingual natural language generation
A method includes preparing a base model using an input model pretrained on at least three languages different from each other and a base vocabulary including words corresponding to two languages among the at least three languages, where the preparing the base model includes constraining the input model to the words included in the base vocabulary; training the base model using a first enhanced training dataset generated from public data, to generate a text summarization model; training the base model using a second enhanced training dataset generated from the first enhanced training dataset, to generate a text generation model; and training the base model using a third enhanced training dataset that is generated using the second enhanced training dataset and the text summarization model, to generate a next sentence generation model.
1 . A computer-implemented method comprising:
preparing a base model using an input model pretrained on at least three languages different from each other and a base vocabulary including words corresponding to two languages among the at least three languages, wherein the preparing the base model comprises constraining the input model to the words included in the base vocabulary;
training the base model using a first enhanced training dataset generated from public data, to generate a text summarization model;
training the base model using a second enhanced training dataset generated from the first enhanced training dataset, to generate a text generation model; and
training the base model using a third enhanced training dataset that is generated using the second enhanced training dataset and the text summarization model, to generate a next sentence generation model,
wherein the generated text generation model is trained to, based on an input of one or more first keywords, output text including at least one first keyword among the one or more first keywords,
wherein the generated text summarization model is trained to, based on an input of the text, output a text summary, and
wherein the generated next sentence generation model is trained to, based on an input of one or more second keywords and the text summary, output a next sentence that is appendable to the text generated by the text generation model and includes at least one second keyword among the one or more second keywords.
2 . The computer-implemented method of claim 1 , wherein each of the text generation model, the text summarization model, and the next sentence generation model is a bilingual model that is trained on the two languages including a target language and English, and configured to output predictions based on an input provided in the target language, English, or a mixed language in which the target language and English are intermixed.
3 . The computer-implemented method of claim 1 , wherein the input model is a transformer-based model.
4 . The computer-implemented method of claim 1 , wherein the input model is an mT5 model.
5 . The computer-implemented method of claim 4 , wherein:
the base model is a modified mT5 model, and
in the base model, a vocabulary of the mT5 model is restricted to a first number of words in a target language and a second number of words in English.
6 . The computer-implemented method of claim 1 , wherein the first enhanced training dataset comprises article-summary pairs, and for each article-summary pair, an article serves as an input training datapoint and a corresponding summary serves as a given output,
the second enhanced training dataset comprises keyword-text pairs, and, for each keyword-text pair, one or more keywords serve as an input training datapoint and a corresponding text serves as a given output, and
the third enhanced training dataset comprises summary-keywords-next sentence triplets, and, for each summary-keywords-next sentence triplet, a summary and keywords serve as an input training datapoint and a corresponding next sentence serves as a given output.
7 . The computer-implemented method of claim 1 , further comprising:
generating a plurality of refined models by training the text generation model and the next sentence generation model using a plurality of refined training datasets generated using the text summarization model and private data.
8 . The computer-implemented method of claim 7 , wherein:
the plurality of refined models comprises a refined text generation model and a refined next sentence generation model,
the generating the plurality of refined models further comprises:
training the text generation model using a first refined training dataset among the plurality of refined training datasets, to generate the refined text generation model, wherein the first refined training dataset is generated based on the private data and includes fake values given to entity values included in the private data, and
training the next sentence generation model using a second refined training dataset among the plurality of refined training datasets, to generate the refined next sentence generation model, wherein the second refined training dataset is generated using the first refined training dataset and the text summarization model,
the refined text generation model is trained to, based on an input of one or more first entity values, output text including at least one first entity value among the one or more first entity values, and
the refined next sentence generation model is trained to, based on an input of one or more second entity values and a primary text summary generated by the text summarization model based on the text output by the refined text generation model, output a next sentence that is appendable to the text output by the refined text generation model and includes at least one second entity value among the one or more second entity values.
9 . A system comprising:
one or more processors;
a memory that is coupled to the one or more processors and stores one or more instructions that, when executed by the one or more processors, cause the one or more processors to perform a method including:
preparing a base model using an input model pretrained on at least three languages different from each other and a base vocabulary including words corresponding to two languages among the at least three languages, wherein the preparing the base model includes constraining the input model to the words included in the base vocabulary;
training the base model using a first enhanced training dataset generated from public data, to generate a text summarization model;
training the base model using a second enhanced training dataset generated from the first enhanced training dataset, to generate a text generation model; and
training the base model using a third enhanced training dataset that is generated using the second enhanced training dataset and the text summarization model, to generate a next sentence generation model,
wherein the generated text generation model is trained to, based on an input of one or more first keywords, output text including at least one first keyword among the one or more first keywords,
wherein the generated text summarization model is trained to, based on an input of the text, output a text summary, and
wherein the generated next sentence generation model is trained to, based on an input of one or more second keywords and the text summary, output a next sentence that is appendable to the text generated by the text generation model and includes at least one second keyword among the one or more second keywords.
10 . The system of claim 9 , wherein each of the text generation model, the text summarization model, and the next sentence generation model is a bilingual model that is trained on the two languages including a target language and English, and configured to output predictions based on an input provided in the target language, English, or a mixed language in which the target language and English are intermixed.
11 . The system of claim 9 , wherein:
the base model is a modified mT5 model, and
in the base model, a vocabulary of an mT5 model is restricted to a first number of words in a target language and a second number of words in English.
12 . The system of claim 9 , wherein the first enhanced training dataset comprises article-summary pairs, and for each article-summary pair, an article serves as an input training datapoint and a corresponding summary serves as a given output,
the second enhanced training dataset comprises keyword-text pairs, and, for each keyword-text pair, one or more keywords serve as an input training datapoint and a corresponding text serves as a given output, and
the third enhanced training dataset comprises summary-keywords-next sentence triplets, and, for each summary-keywords-next sentence triplet, a summary and keywords serve as an input training datapoint and a corresponding next sentence serves as a given output.
13 . The system of claim 9 , wherein the method further includes:
generating a plurality of refined models by training the text generation model and the next sentence generation model using a plurality of refined training datasets generated using the text summarization model and private data.
14 . The system of claim 13 , wherein:
the plurality of refined models comprises a refined text generation model and a refined next sentence generation model,
the generating the plurality of refined models further includes:
training the text generation model using a first refined training dataset among the plurality of refined training datasets, to generate the refined text generation model, wherein the first refined training dataset is generated based on the private data and includes fake values given to entity values included in the private data, and
training the next sentence generation model using a second refined training dataset among the plurality of refined training datasets, to generate the refined next sentence generation model, wherein the second refined training dataset is generated using the first refined training dataset and the text summarization model,
the refined text generation model is trained to, based on an input of one or more first entity values, output text including at least one first entity value among the one or more first entity values, and
the refined next sentence generation model is trained to, based on an input of one or more second entity values and a primary text summary generated by the text summarization model based on the text output by the refined text generation model, output a next sentence that is appendable to the text output by the refined text generation model and includes at least one second entity value among the one or more second entity values.
15 . A non-transitory computer-readable memory storing one or more instructions that, when executed by one or more processors, cause the one or more processors to perform a method including:
preparing a base model using an input model pretrained on at least three languages different from each other and a base vocabulary including words corresponding to two languages among the at least three languages, wherein the preparing the base model includes constraining the input model to the words included in the base vocabulary;
training the base model using a first enhanced training dataset generated from public data, to generate a text summarization model;
training the base model using a second enhanced training dataset generated from the first enhanced training dataset, to generate a text generation model; and
training the base model using a third enhanced training dataset that is generated using the second enhanced training dataset and the text summarization model, to generate a next sentence generation model,
wherein the generated text generation model is trained to, based on an input of one or more first keywords, output text including at least one first keyword among the one or more first keywords,
wherein the generated text summarization model is trained to, based on an input of the text, output a text summary, and
wherein the generated next sentence generation model is trained to, based on an input of one or more second keywords and the text summary, output a next sentence that is appendable to the text generated by the text generation model and includes at least one second keyword among the one or more second keywords.
16 . The non-transitory computer-readable memory of claim 15 , wherein each of the text generation model, the text summarization model, and the next sentence generation model is a bilingual model that is trained on the two languages including a target language and English, and configured to output predictions based on an input provided in the target language, English, or a mixed language in which the target language and English are intermixed.
17 . The non-transitory computer-readable memory of claim 15 , wherein:
the base model is a modified mT5 model, and
in the base model, a vocabulary of an mT5 model is restricted to a first number of words in a target language and a second number of words in English.
18 . The non-transitory computer-readable memory of claim 15 , wherein the first enhanced training dataset comprises article-summary pairs, and for each article-summary pair, an article serves as an input training datapoint and a corresponding summary serves as a given output,
the second enhanced training dataset comprises keyword-text pairs, and, for each keyword-text pair, one or more keywords serve as an input training datapoint and a corresponding text serves as a given output, and
the third enhanced training dataset comprises summary-keywords-next sentence triplets, and, for each summary-keywords-next sentence triplet, a summary and keywords serve as an input training datapoint and a corresponding next sentence serves as a given output.
19 . The non-transitory computer-readable memory of claim 15 , wherein the method further includes:
generating a plurality of refined models by training the text generation model and the next sentence generation model using a plurality of refined training datasets generated using the text summarization model and private data.
20 . The non-transitory computer-readable memory of claim 19 , wherein:
the plurality of refined models comprises a refined text generation model and a refined next sentence generation model,
the generating the plurality of refined models further includes:
training the text generation model using a first refined training dataset among the plurality of refined training datasets, to generate the refined text generation model, wherein the first refined training dataset is generated based on the private data and includes fake values given to entity values included in the private data, and
training the next sentence generation model using a second refined training dataset among the plurality of refined training datasets, to generate the refined next sentence generation model, wherein the second refined training dataset is generated using the first refined training dataset and the text summarization model,
the refined text generation model is trained to, based on an input of one or more first entity values, output text including at least one first entity value among the one or more first entity values, and
the refined next sentence generation model is trained to, based on an input of one or more second entity values and a primary text summary generated by the text summarization model based on the text output by the refined text generation model, output a next sentence that is appendable to the text output by the refined text generation model and includes at least one second entity value among the one or more second entity values.