Implementing Rendered Fabrication Results with Computer Numerically Controlled Machines Based on Natural Language Descriptions of Desired Fabrication Results

Patent metadata

Field	Value
Publication number	US-2024176321-A1
Application number	US-202318525827-A
Country	US
Kind code	A1
Filing date	Nov 30, 2023
Priority date	Nov 30, 2022
Publication date	May 30, 2024
Grant date	—

Abstract

Embodiments include one or more computing systems configured to perform functions comprising: (i) receiving a natural language description of a desired fabrication result; (ii) causing a software engine to generate machine-created rendering instructions based on the natural language description of the desired fabrication result; and (iii) executing, by a laser CNC machine, the machine-created rendering instructions to generate a rendered fabrication result.

First claim

1 - 20 . (canceled)

21 . Tangible, non-transitory computer-readable media comprising program instructions, wherein the program instructions, when executed by one or more processors, cause a computing system to perform functions comprising:
receiving an input comprising a natural language description of a desired fabrication result; generating a prompt for a generative artificial intelligence (AI) model based on the input, wherein the prompt comprises (i) at least a portion of the natural language description of the desired fabrication result and (ii) data representing a continuous vector in a latent space of the generative AI model; causing the generative AI model to generate a design based on the prompt; and after receiving the design from the generative AI model based on the prompt, generating rendering instructions based on the design, wherein the rendering instructions, when input to a laser CNC machine, cause the laser CNC machine to generate a rendered fabrication result based on the rendering instructions.

22 . The tangible, non-transitory computer-readable media of claim 21 , wherein the generative AI model comprises a special-purpose generative AI model that has been trained to generate designs compatible for execution by a laser CNC machine.

23 . The tangible, non-transitory computer-readable media of claim 21 , wherein the generative AI model comprises a general generative AI model that has been fine-tuned to generate designs compatible for execution by a laser CNC machine.

24 . The tangible, non-transitory computer-readable media of claim 21 , wherein the functions further comprise:
causing the laser CNC machine to generate the rendered fabrication result based on the rendering instructions.

25 . The tangible, non-transitory computer-readable media of claim 21 , wherein the input comprising the natural language description of the desired fabrication result further comprises one or more of (i) text data, (ii) image data, (iii) audio data, (iii) video data, (iv) streaming data, (v) sensor data, or (vi) Computer-Aided Design (CAD) data.

26 . The tangible, non-transitory computer-readable media of claim 21 , wherein the prompt further comprises one or more rendering parameters, wherein the one or more rendering parameters comprise at least one of (i) a type of laser CNC machine or (ii) an aspect of a material upon which or from which the desired fabrication result is to be generated.

27 . The tangible, non-transitory computer-readable media of claim 21 , wherein the functions further comprise:
causing generation of a graphical representation of a prospective fabrication result corresponding to the design received from the generative AI model; causing the graphical representation of the prospective fabrication result to be displayed via a graphical user interface; and after receiving an approval of the graphical representation of the prospective fabrication result, generating the rendering instructions based on the design.

28 . The tangible, non-transitory computer-readable media of claim 21 , wherein the functions further comprise:
causing the generative AI model to generate a plurality of designs based on the prompt; for each design of the plurality of designs generated by the generative AI model, causing generation of a graphical representation of a prospective fabrication result corresponding to the design; for each prospective fabrication result corresponding to each design of the plurality of designs generated by the generative AI model, displaying the prospective fabrication result via a graphical user interface; after receiving a selection of one prospective fabrication result, generating rendering instructions based on the design corresponding to the selected prospective fabrication result.

29 . The tangible, non-transitory computer-readable media of claim 21 , wherein generating rendering instructions based on the design comprises:
generating initial rendering instructions; determining whether the initial rendering instructions comprise one or more aspects that will result in a subpar rendered fabrication result; and for an individual aspect of the initial rendering instructions that will result in an unsatisfactory rendered fabrication result, remove or modify the individual aspect of the initial rendering instructions to create the rendering instructions used to generate the rendered fabrication result.

30 . The tangible, non-transitory computer-readable media of claim 21 , wherein the laser CNC machine comprises the computing system.

31 . A method performed by a computing system, the method comprising:
receiving an input comprising a natural language description of a desired fabrication result; generating a prompt for a generative artificial intelligence (AI) model based on the input, wherein the prompt comprises (i) at least a portion of the natural language description of the desired fabrication result and (ii) data representing a continuous vector in a latent space of the generative AI model; causing the generative AI model to generate a design based on the prompt; and after receiving the design from the generative AI model based on the prompt, generating rendering instructions based on the design, wherein the rendering instructions, when input to a laser CNC machine, cause the laser CNC machine to generate a rendered fabrication result based on the rendering instructions.

32 . The method of claim 31 , wherein the generative AI model comprises a special-purpose generative AI model that has been trained to generate designs compatible for execution by a laser CNC machine.

33 . The method of claim 31 , wherein the generative AI model comprises a general generative AI model that has been fine-tuned to generate designs compatible for execution by a laser CNC machine.

34 . The method of claim 31 , further comprising:
causing the laser CNC machine to generate the rendered fabrication result based on the rendering instructions.

35 . The method of claim 31 , wherein the input comprising the natural language description of the desired fabrication result further comprises one or more of (i) text data, (ii) image data, (iii) audio data, (iii) video data, (iv) streaming data, (v) sensor data, or (vi) Computer-Aided Design (CAD) data.

36 . The method of claim 31 , wherein the prompt further comprises one or more rendering parameters, wherein the one or more rendering parameters comprise at least one of (i) a type of laser CNC machine or (ii) an aspect of a material upon which or from which the desired fabrication result is to be generated.

37 . The method of claim 31 , further comprising:
causing generation of a graphical representation of a prospective fabrication result corresponding to the design received from the generative AI model; causing the graphical representation of the prospective fabrication result to be displayed via a graphical user interface; and after receiving an approval of the graphical representation of the prospective fabrication result, generating the rendering instructions based on the design.

38 . The method of claim 31 , further comprising:
causing the generative AI model to generate a plurality of designs based on the prompt; for each design of the plurality of designs generated by the generative AI model, causing generation of a graphical representation of a prospective fabrication result corresponding to the design; for each prospective fabrication result corresponding to each design of the plurality of designs generated by the generative AI model,

Assignees

• Glowforge Inc

Inventors

• Shapiro Daniel
• Natkin Michael
• Park Jonathan
• Shen Jack
• Quan Adrian

Classifications

• B33Y50/00

  Data acquisition or data processing for additive manufacturing · CPC title

• G06F40/30

  Semantic analysis · CPC title

• G05B19/182 Primary

  characterised by the machine tool function, e.g. thread cutting, cam making, tool direction control (G05B19/21 - G05B19/40 take precedence) · CPC title

• G06T19/00

  Manipulating three-dimensional [3D] models or images for computer graphics · CPC title

• G06T2200/24

  involving graphical user interfaces [GUIs] · CPC title