US 12,117,805 B2
System and method for generating industrial processes in a computing environment
Muhammad Zeeshan Zia, Sammamish, WA (US); Quoc-Huy Tran, Redmond, WA (US); and Andrey Konin, Redmond, WA (US)
Assigned to Retrocausal, Inc.
Filed by RETROCAUSAL, INC., Redmond, WA (US)
Filed on Jul. 11, 2023, as Appl. No. 18/350,004.
Claims priority of provisional application 63/445,187, filed on Feb. 13, 2023.
Prior Publication US 2024/0272618 A1, Aug. 15, 2024
Int. Cl. G05B 19/4155 (2006.01)
CPC G05B 19/4155 (2013.01) [G05B 2219/31449 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A computer-implemented system for generating one or more machine learning based (ML-based) industrial processes in a computing environment the computer-implemented system comprising:
one or more hardware processors:
a memory coupled to the one or more hardware processors, wherein the memory comprises a plurality of subsystems in a form of programmable instructions executable by the one or more hardware processors, wherein the plurality of subsystems comprises:
an input receiving subsystem configured to receive one or more multimedia inputs from one or more users;
a semantic determining subsystem configured to analyze the received one or more multimedia inputs to determine semantics associated with the one or more multimedia inputs, using at least one of: natural language processing techniques, symbolic processing techniques, and deep learning techniques, wherein the semantics comprises at least one of: contexts, nuances, linguistics, and modalities;
a process determining subsystem configured to determine one or more process specifications and one or more descriptions in one or more industrial processes, using machine learning (ML) models, wherein the ML models comprise transformer models, autoencoders, generative adversarial networks, and autoregressive models, wherein the one or more process specifications comprises at least one of: a work cell layout for an assembly operation, a first standard operating procedure for the assembly operation, and a computer language description of the one or more industrial processes, wherein the one or more descriptions comprises a natural language description of the one or more industrial processes;
a machine learning (ML) model training subsystem configured to train the ML models on a library of the one or more process specifications and the one or more descriptions;
a generative modeling subsystem configured to perform a generative modeling of the ML models using at least one of: optimization algorithms, machine learning model architectures, and an objective function:
an output tuning subsystem configured to tune weights and parameters of the ML models to return output corresponding to the one or more multimedia inputs, during the training of the ML models;
a process type combining subsystem configured to combine one or more process specifications and the one or more descriptions;
a process generating subsystem configured to generate the one or more ML-based industrial processes corresponding to the one or more multimedia inputs, based on combining the one or more process specifications and the one or more descriptions; and
a process outputting subsystem configured to output the generated one or more ML-based industrial processes, on one of: a display of a user device and one or more external devices, and wherein the process outputting subsystem comprises a process description outputting subsystem configured to output a description of the one or more ML-based industrial processes with a three-dimensional (3D) workstation layout and second standard operating procedures, wherein the process description outputting subsystem is configured to output the description of the one or more ML-based industrial processes as one of: augmented reality (AR) and virtual reality (VR) representations.