Generative AI industrial digital technology transfer
A digital technology transfer system transforms technology transfer documents to a set of digitized manufacturing procedures and operations documentation. The system can leverage generative artificial intelligence (AI) and associated trained custom models to transform a technology transfer document to a hierarchical structured model representing a package, or product to be manufactured, and the process for manufacturing the product. The resulting package model can be integrated into a larger model representing an ecosystem of manufacturing entities and plant facilities by assigning steps of the manufacturing process to one or more selected production lines. To reduce dependency on custom-built parsers for each type of document format, the system leverages generative artificial intelligence (AI) and associated trained models to extract and organize document content, and to map the document content to appropriate target data structures that can be consumed by control systems and devices.
1 . A system, comprising:
a processor, operatively coupled to a memory, that executes executable components stored on the memory, the executable components comprising:
a user interface component configured to receive, from a technology owner, a technology transfer document containing information about a product to be manufactured and describing a manufacturing process for manufacturing the product;
a conversion component configured to
extract content from the technology transfer document as content modules,
extract, as a subset of the content modules, table data from a table contained in the technology transfer document, and
generate a package model comprising a hierarchically structured organization of the content modules representing content sections of the technology transfer document; and
a generative artificial intelligence (AI) component configured to formulate and submit prompts to a generative AI model based on the content of the technology transfer document and industrial knowledge encoded in one or more custom models,
wherein the generative AI component is configured to formulate the prompts to prompt the generative AI model for responses that provide instructions to the conversion component for at least one of extracting the content from the technology transfer document, extracting the table data from the table, or generating the package model.
2 . The system of claim 1 , wherein the one or more custom models are trained using training data comprising at least one of content index data defining a range of content to be extracted from the technology transfer document, semantic rules that define global or customer-specific parsing rules for normalizing the table data, or knowledge of industrial applications, information defining industrial standards.
3 . The system of claim 1 , wherein the conversion component is configured to organize the content modules in the package model according to hierarchical levels comprising one or more of a process level representing a manufacturing process, a stage level representing stages of the manufacturing process, a step level representing steps of the stages, or a parameter level representing control parameter values associated with the steps.
4 . The system of claim 3 , wherein the generative AI component is further configured to determine equipment requirements for the manufacturing process defined in the package model, and to select a manufacturing entity, from among multiple candidate manufacturing entities, to be assigned the manufacturing process based on a determination that the manufacturing entity has equipment capability that meets or exceeds the equipment requirements.
5 . The system of claim 4 , wherein the generative AI component is configured to determine the equipment requirements based in part on responses returned by the generative AI model in response to prompts formulated by the generative AI component based on content of the package model.
6 . The system of claim 4 , wherein the executable components further comprise an export component configured to translate information about the manufacturing process contained in the package model to control configuration data and to export the control configuration data to an industrial device or system associated with the manufacturing entity, wherein the control configuration data configures the industrial device or system to execute a portion of the manufacturing process.
7 . The system of claim 6 , wherein the generative AI component is configured to formulate prompts that prompt the generative AI model for responses that provide instructions to the export component for translating the information about the manufacturing process contained in the package model to the control configuration data in a format supported by the industrial device or system.
8 . The system of claim 1 , wherein the content modules comprise modules of different types of content, the different types of content comprising at least one of text, images, tables, or flow diagrams.
9 . The system of claim 1 , wherein the technology transfer document is imported as at least one of a portable document format file, a word processing file, or an image file.
10 . The system of claim 1 , wherein the executable components further comprise a user interface configured to
render the package model as a browsable navigation tree on a user interface, and
in response to selection, via interaction with the user interface, of a node of the navigation tree representing a section of the technology transfer document, render content associated with the section of the technology transfer document on the user interface.
11 . A method, comprising:
receiving, by a system comprising a processor, a technology transfer document from a technology owner, the technology transfer document containing information about a product to be manufactured and describing a manufacturing process for manufacturing the product;
extracting, by the system, content from the technology transfer document as content modules;
extracting, by the system as a subset of the content modules, table data from a table contained in the technology transfer document; and
generating, by the system, a package model comprising a hierarchically structured organization of the content modules representing content sections of the technology transfer document,
wherein at least one of the extracting of the content or the generating of the package model comprises formulating and submitting prompts to a generative AI model based on the content of the technology transfer document and industrial knowledge encoded in one or more custom models, wherein the prompts are designed to prompt the generative AI model for responses that provide instructions for performing the at least one of the extracting of the content or the generating of the package model.
12 . The method of claim 11 , further comprising training the one or more custom models using training data comprising at least one of content index data defining a range of content to be extracted from the technology transfer document, semantic rules that define global or customer-specific parsing rules for normalizing the table data, or knowledge of industrial applications, information defining industrial standards.
13 . The method of claim 11 , wherein the generating of the package model comprises organizing the content modules in the package model according to hierarchical levels comprising one or more of a process level representing a manufacturing process, a stage level representing stages of the manufacturing process, a step level representing steps of the stages, or a parameter level representing control parameter values associated with the steps.
14 . The method of claim 13 , further comprising:
determining, by the system, equipment requirements for the manufacturing process defined in the package model; and
selecting, by the system, a manufacturing entity, from among multiple candidate manufacturing entities, to be assigned the manufacturing process based on a determination that the manufacturing entity has equipment capability that meets or exceeds the equipment requirements.
15 . The method of claim 14 , wherein at least one of the determining of the equipment requirements or the selecting of the manufacturing entity comprises is based in part on responses returned by the generative AI model in response to prompts formulated by the system based on content of the package model.
16 . The method of claim 14 , further comprising:
translating, by the system, information about the manufacturing process contained in the package model to control configuration data; and
exporting, by the system, the control configuration data to an industrial device or system associated with the manufacturing entity, wherein the control configuration data configures the industrial device or system to execute a portion of the manufacturing process.
17 . The method of claim 16 , further comprising formulating prompts that prompt the generative AI model for responses that provide instructions for performing the translating of the information about the manufacturing process to the control configuration data in a format supported by the industrial device or system.
18 . The method of claim 11 , wherein the content modules comprise modules of different types of content, the different types of content comprising at least one of text, images, tables, or flow diagrams.
19 . A non-transitory computer-readable medium having stored thereon instructions that, in response to execution, cause a technology transfer system comprising a processor to perform operations, the operations comprising:
importing a technology transfer document associated with a technology owner, the technology transfer document comprising information describing a product to be manufactured and a manufacturing process for manufacturing the product;
performing a first parsing process on the technology transfer document that identifies and extracts content from the technology transfer document as content modules;
performing a second parsing process on the technology transfer document that extracts, as a subset of the content modules, table data from a table contained in the technology transfer document; and
generating a package model comprising a hierarchically structured organization of the content modules representing content sections of the technology transfer document,
wherein
at least one of the performing of the first parsing process, the performing of the second parsing process, or the generating of the package model comprises formulating and submitting prompts to a generative AI model based on the content of the technology transfer document and industrial knowledge encoded in one or more custom models, and
the prompts are designed to prompt the generative AI model for responses that provide instructions for the at least one of the performing of the first parsing process, the performing of the second parsing process, or the generating of the package model.
20 . The non-transitory computer-readable medium of claim 19 , further comprising training the one or more custom models using training data comprising at least one of content index data defining a range of content to be extracted from the technology transfer document, semantic rules that define global or customer-specific parsing rules for normalizing the table data, or knowledge of industrial applications, information defining industrial standards.