Schemaless systems and methods for automatically building and utilizing a chatbot knowledge base or the like
Schemaless systems and methods for automatically building and utilizing a chatbot knowledge base or the like. Textual analytics tools, such as natural language processing (NLP) and/or deep learning engines, are used to analyze the knowledge base and uncover and highlight gaps, which are turned into topics. Predetermined answers to these queries can then be prepared manually, or by an artificial intelligence (AI) algorithm with alternative database visibility. In this manner, new question-and-answer (Q&A) pairs are generated by the systems and methods in an automated manner, for later use by a chatbot, coaching system, or the like. These processes are iterative. Advantageously, transferring the processes from manual control to automated control greatly conserves resources.
1. A system for automatically building a knowledge base, comprising:
memory storing question-and-answer detection instructions forming a textual analytics tool executed by a processor and operable for parsing a query to a platform utilizing an existing knowledge base to identify a missing question-and-answer pair associated with the existing knowledge base;
the memory storing answer formulation instructions executed by the processor and operable for manually or automatically generating an answer to the query to form a potential question-and-answer pair;
the memory storing question-and-answer validation instructions executed by the processor and operable for manually or automatically validating the formulated potential question-and-answer pair; and
the memory storing question-and-answer substitution instructions executed by the processor and operable for updating the existing knowledge base with the validated formulated potential question-and-answer pair to form a new knowledge base;
wherein the potential question-and-answer pair associated with the query and question-and-answer pairs associated with the existing knowledge base are vector encoded by a deep neural network (DNN) and compared for semantic similarity distance to determine if the existing knowledge base is expanded to form the new knowledge base, wherein the DNN is trained with training data utilizing a text clustering algorithm to discover hidden topics and provide a taxonomy that informs data inclusion in the new knowledge base; and
wherein the platform is operable for receiving another query and responding utilizing the new knowledge base.
2. The system of claim 1 , wherein the textual analytics tool comprises one or more of a natural language processing engine and a deep learning engine.
3. The system of claim 1 , wherein the textual analytics tool is operable for generating one or more topics associated with the query.
4. The system of claim 1 , wherein manually generating the answer to the query comprises presenting the query to a user and receiving the answer from the user.
5. The system of claim 1 , wherein automatically generating the answer to the query comprises transmitting the query to an artificial intelligence engine and receiving the answer from the artificial engine.
6. The system of claim 1 , wherein manually validating the formulated potential question-and-answer pair comprises presenting the formulated potential question-and-answer pair to a user for approval.
7. The system of claim 1 , wherein automatically validating the formulated potential question-and-answer pair comprises presenting the formulated potential question-and-answer pair to an artificial intelligence engine for approval.
8. The system of claim 1 , wherein the platform comprises one of a chatbot and a coaching system.
9. The system of claim 1 , wherein the existing knowledge base is updated iteratively to form the new knowledge base.
10. A method for automatically building a knowledge base, comprising:
parsing a query to a platform utilizing an existing knowledge base to identify a missing question-and-answer pair associated with the existing knowledge base;
manually or automatically generating an answer to the query to form a potential question-and-answer pair;
manually or automatically validating the formulated potential question-and-answer pair; and
updating the existing knowledge base with the validated formulated potential question-and-answer pair to form a new knowledge base;
wherein the potential question-and-answer pair associated with the query and question-and-answer pairs associated with the existing knowledge base are vector encoded by a deep neural network (DNN) and compared for semantic similarity distance to determine if the existing knowledge base is expanded to form the new knowledge base, wherein the DNN is trained with training data utilizing a text clustering algorithm to discover hidden topics and provide a taxonomy that informs data inclusion in the new knowledge base; and
wherein the platform is operable for receiving another query and responding utilizing the new knowledge base.
11. The method of claim 10 , wherein parsing the query comprises parsing the query using one or more of a natural language processing engine and a deep learning engine.
12. The method of claim 10 , wherein parsing the query comprises generating one or more topics associated with the query.
13. The method of claim 10 , wherein manually generating the answer to the query comprises presenting the query to a user and receiving the answer from the user.
14. The method of claim 10 , wherein automatically generating the answer to the query comprises transmitting the query to an artificial intelligence engine and receiving the answer from the artificial engine.
15. The method of claim 10 , wherein manually validating the formulated potential question-and-answer pair comprises presenting the formulated potential question-and-answer pair to a user for approval.
16. The method of claim 10 , wherein automatically validating the formulated potential question-and-answer pair comprises presenting the formulated potential question-and-answer pair to an artificial intelligence engine for approval.
17. The method of claim 10 , wherein the platform comprises one of a chatbot and a coaching system.
18. The method of claim 10 , wherein the existing knowledge base is updated iteratively to form the new knowledge base.
19. A method for using a knowledge base, comprising:
receiving, at a platform, a query from a user; and
accessing a knowledge base comprising a plurality of question-and-answer pairs to locate a question-and-answer pair that comprises an answer to the query from the user;
wherein the knowledge base is formed by:
parsing a query to the platform utilizing a prior knowledge base to identify a missing question-and-answer pair associated with the prior knowledge base;
manually or automatically generating an answer to the query to form a potential question-and-answer pair;
manually or automatically validating the formulated potential question-and-answer pair; and
updating the existing knowledge base with the validated formulated potential question-and-answer pair to form the knowledge base;
wherein the potential question-and-answer pair associated with the query and question-and-answer pairs associated with the prior knowledge base are vector encoded by a deep neural network (DNN) and compared for semantic similarity distance to determine if the prior knowledge base is expanded to form the knowledge base, wherein the DNN is trained with training data utilizing a text clustering algorithm to discover hidden topics and provide a taxonomy that informs data inclusion in the new knowledge base.
20. The method of claim 19 , wherein parsing the query comprises parsing the query using one or more of a natural language processing engine and a deep learning engine.
21. The method of claim 19 , wherein the platform comprises one of a chatbot and a coaching system.