View source for 18064230. Building Bots from Raw Logs and Computing Coverage of Business Logic Graph simplified abstract (Google LLC)

=Building Bots from Raw Logs and Computing Coverage of Business Logic Graph=


==Organization Name==

[[:Category:Google LLC|Google LLC]]


[[Category:Google LLC]]

==Inventor(s)==

[[:Category:Joseph Lange of Zurich (CH)|Joseph Lange of Zurich (CH)]][[Category:Joseph Lange of Zurich (CH)]]

[[:Category:Henry Scott Dlhopolsky of Zurich (CH)|Henry Scott Dlhopolsky of Zurich (CH)]][[Category:Henry Scott Dlhopolsky of Zurich (CH)]]

[[:Category:Vladimir Vuskovic of Zollikerberg (CH)|Vladimir Vuskovic of Zollikerberg (CH)]][[Category:Vladimir Vuskovic of Zollikerberg (CH)]]

==Building Bots from Raw Logs and Computing Coverage of Business Logic Graph - A simplified explanation of the abstract==


This abstract first appeared for US patent application 18064230 titled 'Building Bots from Raw Logs and Computing Coverage of Business Logic Graph

==Simplified Explanation==
The patent application describes a method for generating training data for a model based on conversations between a customer and an agent. The method involves selecting responses from a logic model to train a machine learning model.

* The method receives a transcript of a conversation between a customer and an agent.
* It selects a response from a logic model based on the agent's input.
* If the selected response is similar enough to the agent's input, it is used to train a machine learning model.

==Key Features and Innovation==
* Dynamic generation of training data based on real conversations.
* Utilization of a logic model to select responses for training.
* Incorporation of similarity scoring to determine the effectiveness of selected responses.

==Potential Applications==
This technology can be applied in:
* Customer service training programs.
* Chatbot development.
* Call center optimization.

==Problems Solved==
* Efficient generation of training data for machine learning models.
* Improved accuracy in response selection based on real conversations.
* Enhanced performance of AI systems in customer-agent interactions.

==Benefits==
* Streamlined training process for AI models.
* Enhanced customer service experiences.
* Increased efficiency in agent training programs.

==Commercial Applications==
* "Dynamic Training Data Generation for AI Customer Service Models: Market Implications"

==Prior Art==
There may be prior art related to machine learning models trained on conversation transcripts and logic models for response selection.

==Frequently Updated Research==
No specific information on frequently updated research related to this technology.

==Questions about Dynamic Training Data Generation==
===How does this method improve the accuracy of AI models in customer service applications?===
This method enhances accuracy by selecting responses based on real conversations, improving the relevance of training data.

===What are the potential challenges in implementing this technology in different industries?===
The challenges may include adapting the logic model to specific industry jargon and conversation nuances.


==Original Abstract Submitted==

A method for dynamically generating training data for a model includes receiving a transcript corresponding to a conversation between a customer and an agent, the transcript comprising a customer input and an agent input. The method includes receiving a logic model including a plurality of responses, each response of the plurality of responses representing a potential reply to the customer input. The method further includes selecting, based on the agent input, a response from the plurality of responses of the logic model. The method includes determining that a similarity score between the selected response and the agent input satisfies a similarity threshold, and, based on determining that the similarity score between the selected response and the agent input satisfies the similarity threshold, training a machine learning model using the customer input and the selected response.



[[Category:G06F40/35]]
[[Category:G06N20/00]]
[[Category:G06Q30/02]]
[[Category:CPC_G06F40/35]]