Machine learning (ML) powered programming code transformations
US11693641B1 · kind B1 · utility
Assignee
Inventors
Key dates
| Filing date | Mar 22, 2022 |
| Grant date | Jul 4, 2023 |
| Priority date | — |
| Expiry date | Mar 22, 2042 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06N20/00
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
A machine learning (ML) based code transformation system that transforms a source programming code developed using a source library for execution on a source platform into remediated code for execution on a target platform is disclosed. Metadata extracted from the source programming code is used to detect the source programming language, source libraries, and the source platform. The metadata also enables modularizing the source programming code based on the functionality and identifying a node from a plurality of nodes in a communication network to execute the various source code modules. A similarity map is generated mapping the source libraries to the target libraries and the source code modules that are incompatible with the target platform are identified and remediated with similar target code modules using the similarity map.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.