Method and system for generating a conversational agent by automatic paraphrase generation based on machine translation
The present teaching relates to generating a conversational agent. In one example, a plurality of input utterances may be received from a developer. A paraphrase model is obtained. The paraphrase model is generated based on machine translation. For each of the plurality of input utterances, one or more paraphrases of the input utterance are generated based on the paraphrase model. For each of the plurality of input utterances, at least one of the one or more paraphrases is selected based on an instruction from the developer to generate selected paraphrases. The conversational agent is generated based on the plurality of input utterances and the selected paraphrases.
1. A method, implemented on a machine having at least one processor, storage, and a communication platform connected to a network for generating a conversational agent, the method comprising:
parsing each of a plurality of input utterances into multiple portions;
obtaining, based on a paraphrase model according to the multiple portions, multiple paraphrases for each of the plurality of input utterances; and
generating a conversational agent based on at least one of the multiple paraphrases selected by a user, wherein the paraphrase model is updated based on the at least one of the multiple paraphrases selected by the user.
2. The method of claim 1 , wherein the selected at least one of the one or more paraphrases matches at least one of one or more intents of the plurality of input utterances.
3. The method of claim 1 , wherein the paraphrase model is trained via machine translation.
4. The method of claim 1 , further comprising:
obtaining aggregated training data based on one or more of stored user-agent dialogs or the selected at least one of the one or more paraphrases; and
training the paraphrase model based on aggregated training data.
5. The method of claim 4 , further comprising:
selecting, based on the aggregated training data, the paraphrase model from a plurality of paraphrase models.
6. The method of claim 4 , further comprising:
generating a plurality of N-grams features from the aggregated training data;
determining, based on the plurality of N-grams features, that the aggregated training data comprises new training data related to the paraphrase model; and
updating the paraphrase model based on the aggregated training data.
7. The method of claim 4 , further comprising:
generating a plurality of N-grams from the aggregated training data;
determining, based on the plurality of N-grams, that the aggregated training data is unrelated to any previously generated paraphrase model; and
generating the paraphrase model based on the aggregated training data.
8. The method of claim 1 , wherein the selected at least one of the one or more paraphrases is semantically equivalent to a corresponding input utterance.
9. A method, implemented on a machine having at least one processor, storage, and a communication platform connected to a network for generating a conversational agent, the method comprising:
parsing each of a plurality of input utterances into multiple portions;
obtaining, according to the multiple portions, multiple paraphrases for each of the plurality of input utterances based on a paraphrase model and an intent estimated based on the input utterance;
receiving, from a user, a user instruction indicative of a selection of at least one of the multiple paraphrases;
generating a conversational agent based on the user instruction; and
updating the paraphrase model based on the selection.
10. The method of claim 9 , wherein the selected at least one of the one or more paraphrases matches at least one of the intents of the plurality of input utterances.
11. The method of claim 9 , wherein the intent is obtained based on an intent model trained via machine learning.
12. The method of claim 11 , further comprising training the intent model based on user interaction data.
13. The method of claim 9 , wherein the selected at least one of the one or more paraphrases is semantically equivalent to a corresponding input utterance.
14. A non-transitory machine-readable medium having information recorded thereon for generating a conversational agent, wherein the information, when read by the machine, effectuate operations comprising:
parsing each of a plurality of input utterances into multiple portions;
obtaining, based on a paraphrase model according to the multiple portions, multiple paraphrases for each of the plurality of input utterances; and
generating a conversational agent based on at least one of the multiple paraphrases selected by a user, wherein the paraphrase model is updated based on the at least one of the multiple paraphrases selected by the user.
15. The medium of claim 14 , wherein the selected at least one of the one or more paraphrases matches at least one of one or more intents of the plurality of input utterances.
16. The medium of claim 14 , wherein the paraphrase model is trained via machine translation.
17. The medium of claim 14 , wherein the operations further comprise:
obtaining aggregated training data based on one or more of stored user-agent dialogs or the selected at least one of the one or more paraphrases; and
training the paraphrase model based on aggregated training data.
18. The medium of claim 17 , wherein the operations further comprise:
selecting, based on the aggregated training data, the paraphrase model from a plurality of paraphrase models.
19. The medium of claim 17 , wherein the operations further comprise:
generating a plurality of N-grams features from the aggregated training data;
determining, based on the plurality of N-grams features, that the aggregated training data comprises new training data related to the paraphrase model; and
updating the paraphrase model based on the aggregated training data.
20. The medium of claim 17 , wherein the operations further comprise:
generating a plurality of N-grams from the aggregated training data;
determining, based on the plurality of N-grams, that the aggregated training data is unrelated to any previously generated paraphrase model; and
generating the paraphrase model based on the aggregated training data.