Hyperparameter tuning for generative artificial intelligence prompt engineering

Patent metadata

Field	Value
Publication number	US-2025383653-A1
Application number	US-202418745607-A
Country	US
Kind code	A1
Filing date	Jun 17, 2024
Priority date	Jun 17, 2024
Publication date	Dec 18, 2025
Grant date	—

Abstract

Disclosed herein are methods and systems that provide hyperparameter tuning for generative artificial intelligence prompt engineering to users prompting a generative artificial intelligence (GAI) model to return an executable logic code output. Where a user inputs a portion, a prompt is received and an iteration of a hyperparameter tuning process is executed. The assistant receives the prompt from the user and generates a number of hyperparameter sets. Each of the hyperparameter sets and a copy of the prompt are used to generate a number of complete prompts, where each complete prompt corresponds to a hyperparameter set. The complete prompts are submitted to a generative artificial intelligence model, which returns a number of responses corresponding to each complete prompt. A user selects a response, and the hyperparameters associated with the choice are the basis of a next hyperparameter tuning process iteration.

First claim

What is claimed is:

1 . A method, comprising:
receiving, via a user interface, at least a portion of a prompt designed for submission to a generative artificial intelligence model, wherein the prompt requests a response in a form of logic code executable by an industrial automation controller; executing an iteration of a hyperparameter tuning process, comprising:
generating, by a hyperparameter tuning module, a plurality of hyperparameter sets, wherein the hyperparameter tuning module generates the plurality of hyperparameter sets using a hyperparameter tuning model trained to generate hyperparameter sets for iterations of the hyperparameter tuning process based at least in part on hyperparameter sets associated with response selections in prior iterations,
generating, by a prompt generation module, a plurality of complete prompts, wherein each complete prompt comprises one of the plurality of hyperparameter sets and the at least the portion of a prompt,
submitting each complete prompt to the generative artificial intelligence model,
receiving a plurality of responses from the generative artificial intelligence model, wherein each response of the plurality of responses corresponds to one of the plurality of complete prompts,
providing, via the user interface, each response of the plurality of responses, and
receiving, via the user interface, a response selection chosen from one of the plurality of responses; and
executing a next iteration of the hyperparameter tuning process based on the response selection.

2 . The method of claim 1 , wherein:
the executing a next iteration comprises repeating the executing of the hyperparameter tuning process based on a continued tuning request received via the user interface.

3 . The method of claim 1 , wherein:
the hyperparameter tuning model is further trained to generate the hyperparameter sets to create a uniform distribution across a default range for each hyperparameter in the hyperparameter sets for an initial iteration of the hyperparameter tuning process.

4 . The method of claim 1 , further comprising:
during entry of the at least the portion of a prompt: sending entered words to a next-word suggestion model trained to generate a suggestion for a next word to be used in the at least the portion of a prompt subsequent to the entered words, wherein:
next-word suggestion model is further trained to suggest the next word for prompts designed to elicit responses from the generative artificial intelligence model in a form of executable logic code in a coding language to be used in an industrial automation process based at least in part on the entered words;
receiving, from the next-word suggestion model, the suggestion for the next word; providing instructions to display, via the user interface, a selectable indication of the suggestion for the next word; and incorporating, in response to a selection of the selectable indication via the user interface, the next word after the entered words in the at least the portion of a prompt.

5 . The method of claim 4 , wherein the next-word suggestion model comprises a first domain-specific machine-learning model of a plurality of domain-specific machine-learning models, wherein each domain-specific machine-learning model is based on a domain associated with a specific coding language of a plurality of coding languages, the method further comprising:
selecting, based on the domain associated with the coding language, the first domain-specific machine-learning model from the plurality of domain-specific machine-learning models.

6 . The method of claim 5 , further comprising:
receiving a storing indication associated with the response selection via the user interface, and in response to the storing indication, saving, to a repository, a hyperparameter set and the complete prompt associated with the response selection.

7 . The method of claim 6 , further comprising:
periodically fine tuning each of the plurality of domain-specific machine-learning models based on data stored in the repository.

8 . The method of claim 1 , wherein:
providing each response comprises providing instructions to display, via the user interface, each hyperparameter set associated with each response.

9 . The method of claim 1 , wherein:
the hyperparameter tuning model is further trained to generate the hyperparameter sets based on an indication of an initial hyperparameter set in an initial iteration of the hyperparameter tuning process; and the indication of the initial hyperparameter set comprises one of:
receiving, via the user interface, selection of a stored hyperparameter set as the initial hyperparameter set; and
receiving, via the user interface, a user-entered hyperparameter set as the initial hyperparameter set.

10 . The method of claim 1 , wherein the user interface comprises a software development environment for developing logic code, the method further comprising:
integrating a response selection from a prior iteration of the hyperparameter tuning process into logic code under development in the user interface.

11 . A system, comprising:
a coordinator configured to:
receive, via a user interface, at least a portion of a prompt designed for submission to a generative artificial intelligence model, wherein the at least portion of a prompt requests a response in a form of logic code executable by an industrial automation controller;
receive, via the user interface, requests for hyperparameter tuning associated with the portion of a prompt; and
coordinate an iteration of a hyperparameter tuning process in response to the requests for hyperparameter tuning, wherein to coordinate the iteration, the coordinator is configured to:
receive, from a hyperparameter tuning module, a plurality of hyperparameter sets in response to requesting the plurality of hyperparameter sets;
receive, from a prompt generation module, a plurality of complete prompts in response to requesting the plurality of complete prompts;
receive a plurality of responses from the generative artificial intelligence model in response to submitting the plurality of complete prompts to the generative artificial intelligence model;
provide the plurality of responses via the user interface; and
receive, via the user interface, a response selection of one of the plurality of responses;

the prompt generation module configured to:
receive the request for the plurality of complete prompts and the plurality of hyperparameter sets from the coordinator;
generate a plurality of complete prompts, wherein each complete prompt comprises one of the plurality of hyperparameter sets and the at least the portion of a prompt; and
provide the plurality of complete prompts to the coordinator; and
the hyperparameter tuning module, further configured to:
receive, from the coordinator, a request for the plurality of hyperparameter sets; and
generate the plurality of hyperparameter sets, wherein the hyperparameter tuning module generates the plurality of hyperparameter sets using a hyperparameter tuning model trained to generate hyperparameter sets for iterations of the hyperparameter tuning process based at least in part on hyperparameter sets associated with response selections in prior iterations; and
provide the plurality of hyperparameter sets to the coordinator.

12 . The system of claim 11 , wherein:
the hyperparameter tuning model is further trained, in response to a request for a plurality of hyperparameter sets, to generate the hyperparameter sets to create a uniform distribution across a default range for each hyperparameter in the hyperparameter sets for an initial iteration of

Assignees

• Rockwell Automation Tech Inc

Inventors

• He Meiling
• Chowdhury Ankan
• Maturana Francisco
• Da Silva Aderiano M
• Chyla Brady M

Classifications

• G05B19/41885 Primary

  characterised by modeling, simulation of the manufacturing system · CPC title