GERAÇÃO DE CONTROLES PARA MODIFICAR CONTEÚDO VISUAL EM SISTEMAS COMPUTACIONAIS COM BASE EM MODELOS GENERATIVOS

By enabling non-linear interaction and user-directed modifications, the technique addresses the limitations of linear generative models in graphic design, enhancing the flexibility and effectiveness of visual content generation.

BR102026002102A2Pending Publication Date: 2026-11-10FIGMA INC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
BR102026002102
Authority / Receiving Office
BR · BR
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-01-31
Filing Date
2026-01-28
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing generative artificial intelligence models operate in a linear fashion, which may not align with the intuitive and iterative nature of graphic design processes, limiting their effectiveness in generating visual content.

Method used

The technique allows for non-deterministic and non-linear interaction with generative models by enabling forward and backward interaction with generated results, allowing users to select and re-select visual features as a starting point for subsequent iterations, and includes control panels for user-directed modifications.

Benefits of technology

This approach enhances the flexibility and intuitiveness of generating visual content by allowing generative models to contextually generate results based on previous inference rounds, improving the generation of visual assets in graphic design applications.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader
Need to check novelty before this filing date? Find Prior Art

Description

1 / 46 Generating controls to modify visual content in computer systems based on generative models. FUNDAMENTALS OF THE INVENTION Technical Field of the Different Sports

[001] The modalities of this disclosure relate, in general, to graphic design tools and, more specifically, to techniques for generating visual content using generative artificial intelligence models. Description of the Related State of the Art

[002] Generative artificial intelligence models can be used to generate a variety of content, such as text content, image content, or similar. To generate content, a generative artificial intelligence model generally ingests an input prompt specifying parameters of the content to be generated, and the generative artificial intelligence model produces content according to the parameters. These generative artificial intelligence models can ingest a text prompt and produce a response in one or more data modalities based on the application for which the generative artificial intelligence models are trained. Language models can generate text responses to input prompts, while diffusion models can generate image outputs from an input prompt.

[003] In general, interaction with generative artificial intelligence models is a linear process in which a generative artificial intelligence model generates an output in response to a first textual input prompt and receives a second input prompt related to the generated output. In this linear process, an output from the generative artificial intelligence model can be iteratively refined until a desired result is achieved. That is, the content generated in a first round of inference generally serves as context for a second round of Petition 870260008329, dated 01 / 28 / 2026, page 18 / 78 2 / 46 inference that iterates over the content generated in the first round of inference (for example, adds additional details to the content generated in the first round of inference, generates content describing a concept included in the content generated in the first round of inference, etc.).

[004] For some applications, successive textual prompts referring to content generated by a generative artificial intelligence model in previous inference rounds may be sufficient to request that a generative artificial intelligence model generate a desired output. However, for other applications, textual prompts fed into a generative artificial intelligence model in a linear fashion may not correspond to the way users perform various tasks. For example, the linear nature in which generative artificial intelligence models operate may not provide the appropriate infrastructure to allow content generation using forward iteration and returns to previously generated content.In the context of graphic design, where the design process flow executed by a user involves creating visual content (e.g., user interface designs) using forward iteration and returns to previously generated designs, generative artificial intelligence models operating on the linear prompt paradigm may not allow for the generation of visual content in a natural and intuitive way.

[005] As the above illustrates, more effective techniques are needed to generate visual content using generative user interface models. SUMMARY OF THE INVENTION

[006] One embodiment of the present disclosure provides techniques for generating candidate visual features using generative artificial intelligence models and a selection of a visual feature based on which the candidate visual features are generated. An exemplary method, in general, includes receiving a suggestion of a visual feature for modification and a natural language input prompt specifying a design objective to Petition 870260008329, dated 01 / 28 / 2026, page 19 / 78 3 / 46 to be applied to the indicated visual resource. The design objective is translated into a plurality of transformations to be applied to the indicated visual resource. A candidate set of visual resources is generated based on the application of the plurality of transformations to the indicated visual resource, and the candidate set of visual resources is presented for selection in a user interface.

[007] One embodiment of the present disclosure provides techniques for refining visual features using generative artificial intelligence models. An exemplary method, in general, includes receiving a suggestion of a visual feature for modification and a natural language input prompt specifying a design goal to be applied to the suggested visual feature. Visual components in the suggested visual feature and properties of the identified visual components to be modified are identified based on the specified design goal. One or more control panels are populated with one or more controls to modify the identified visual components to achieve the specified design goal. An input is received in at least one of the one or more control panels, and the suggested visual feature is modified based on the received input.

[008] A technical advantage of the disclosed techniques over the previous technique is that the disclosed techniques allow users of generative artificial intelligence models to interact with these models in a non-deterministic and non-linear way. Generative artificial intelligence models can allow forward and backward interaction with results generated by these generative artificial intelligence models. The ability to generate content using forward and backward interaction with results generated by a generative artificial intelligence model improves the ability of generative artificial intelligence models to generate content by flexibly selecting the context based on which a generative artificial intelligence model generates results from a round of inference. For example, generative artificial intelligence models can generate content based on context. Petition 870260008329, dated 01 / 28 / 2026, page 20 / 78 4 / 46 of the results from any previous inference round performed by the generative artificial intelligence model may not be linked to the context of the results from the immediately preceding inference round. In the context of graphic design applications, the techniques presented here allow users to interact with generative artificial intelligence models using an interaction paradigm that replicates the forward and backward progression of visual content generation. Furthermore, these generative artificial intelligence models can generate a plurality of candidate visual features during each inference round, allowing the selection (and re-selection) of different visual features as a starting point for generating subsequent sets of candidate visual features. BRIEF DESCRIPTION OF THE DRAWINGS

[009] In order that the manner in which the above-mentioned characteristics of the various embodiments may be understood in detail, a more particular description of the inventive concepts, briefly summarized above, can be obtained by reference to various embodiments, some of which are illustrated in the accompanying drawings. It should be noted, however, that the accompanying drawings illustrate only typical embodiments of the inventive concepts and, therefore, should not be considered limiting in any way, and that other equally effective embodiments exist.

[0010] Figure 1 illustrates a computer system configured to implement one or more aspects of various modalities of the present disclosure.

[0011] Figure 2 illustrates an initial visual resource and an input prompt specifying a design objective for the visual resource, according to some modalities.

[0012] Figure 3 illustrates a candidate set of visual features generated by a generative artificial intelligence model based on an input prompt specifying a design goal for a visual feature, according to some modalities. Petition 870260008329, dated 01 / 28 / 2026, page 21 / 78 5 / 46

[0013] Figure 4 illustrates the generation of different candidate sets of visual features by a generative artificial intelligence model based on the selection of a visual feature from a previous candidate set of visual features, according to some modalities.

[0014] Figure 5 is a flowchart of method steps for generating visual assets based on an input prompt specifying a design objective for a visual asset and selection of the visual asset based on which a generative artificial intelligence model generates the visual assets, according to some modalities.

[0015] Figure 6 illustrates control panels for adjusting the appearance of a visual resource based on an input prompt specifying a design objective for the visual resource, according to some modalities.

[0016] Figure 7 illustrates changes in a visual resource generated based on adjustments to one or more parameters in control panels to adjust the appearance of a visual resource, according to some modalities.

[0017] Figure 8 illustrates changes in a visual resource generated based on adjustments to parameters identified based on an input prompt specifying a design objective for the visual resource, according to some modalities.

[0018] Figure 9 is a flowchart of method steps for modifying visual assets based on changes in parameters identified by a generative artificial intelligence model from an input prompt specifying a design objective for a visual asset, according to some modalities.

[0019] Figure 10 illustrates a networked computer system for implementing an interactive graphical application platform, according to some modalities. DETAILED DESCRIPTION OF THE INVENTION Petition 870260008329, dated 01 / 28 / 2026, page 22 / 78 6 / 46

[0020] In the following description, numerous specific details are set forth to provide a fuller understanding of the various embodiments. However, it will be apparent to a person skilled in the art that the inventive concepts can be practiced without one or more of these specific details.

[0021] Figure 1 illustrates a computational device 100 configured to implement one or more aspects of various embodiments of the present invention. In one embodiment, the computational device 100 includes a desktop computer, a laptop computer, a smartphone, a personal digital assistant (PDA), a tablet computer, or any other type of computational device configured to receive input, process data, and optionally display images, and is suitable for practicing one or more embodiments. The computational device 100 is configured to run a generative model 122 and a graphic design engine 124 that reside in a memory 116.

[0022] It should be noted that the computing device described herein is illustrative and that any other technically feasible configurations are within the scope of this disclosure. For example, multiple instances of the generative model 122 or the graphic design engine 124 could be run on a set of nodes in a distributed and / or cloud computing system to implement the functionality of the computing device 100. In another example, the generative model 122 or the graphic design engine 124 could be run on various sets of hardware, device types, or environments to adapt the generative model 122 or the graphic design engine 124 to different use cases or applications. In a third example, the generative model 122 or the graphic design engine 124 could be run on different computing devices and / or different sets of computing devices.

[0023] In one embodiment, the computing device 100 includes, without limitation, an interconnection (bus) 112 that connects one or Petition 870260008329, dated 01 / 28 / 2026, page 23 / 78 7 / 46 plus processors 102, an input / output (I / O) device interface 104 coupled to one or more input / output (I / O) devices 108, memory 116, storage 114, and a network interface 106. The processor(s) 102 may be any suitable processor implemented as a central processing unit (CPU), a graphics processing unit (GPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), an artificial intelligence (AI) accelerator, any other type of processing unit, or a combination of different processing units, such as a CPU configured to operate in conjunction with a GPU. In general, the processor(s) 102 may be any technically feasible hardware unit capable of processing data and / or executing software applications.Furthermore, in the context of this disclosure, the computing elements shown in computing device 100 may correspond to a physical computing system (for example, a system in a data center) or may be a virtual computing instance running within a computing cloud.

[0024] I / O devices 108 include devices capable of providing input, such as a keyboard, a mouse, a touch screen, a microphone, and so forth, as well as devices capable of providing output, such as a display device or loudspeaker. Additionally, I / O devices 108 may include devices capable of receiving input and providing output, such as a touch screen, a Universal Serial Bus (USB) port, and so forth. I / O devices 108 may be configured to receive various types of input from an end user (e.g., a designer) from the computing device 100, and also to provide various types of output to the end user of the computing device 100, such as displayed digital images or digital videos or text. In some embodiments, one or more of the I / O devices 108 are configured to couple the computing device 100 to a network 110.

[0025] The 110 network is any technically feasible type of Petition 870260008329, dated 01 / 28 / 2026, page 24 / 78 8 / 46 A communications network that allows data to be exchanged between the computing device 100 and external entities or devices, such as a web server or another networked computing device. For example, the network 110 may include a wide area network (WAN), a local area network (LAN), a wireless network (Wi-Fi), and / or the Internet, among others.

[0026] Storage 114 includes non-volatile storage for applications and data, and may include fixed or removable disk drives, flash memory devices, and CD-ROM, DVD-ROM, Blu-ray, HDDVD, or other magnetic, optical, or solid-state storage devices. The generative model 122 and the graphic design engine 124 may be stored in storage 114 and loaded into memory 116 when executed.

[0027] Memory 116 includes a random access memory (RAM) module, a flash memory unit, or any other type of memory unit or combination thereof. Processor(s) 102, I / O device interface 104, and network interface 106 are configured to read data from and write data to memory 116. Memory 116 includes various software programs that can be executed by processor(s) 102 and application data associated with said software programs, including the generative model 122 or the graphic design engine 124. Example of Generating Candidate Visual Resources Using Generative Artificial Intelligence Models

[0028] In graphic design software, such as that implemented by the 124 graphic design engine illustrated in Figure 1, visual assets, such as user interface components in a user interface design in the 124 graphic design engine, can be rendered and projected. A visual asset, as used herein, can be a group of visual components rendered in a user interface. Visual assets can be vector assets defined in terms of mathematical relationships between different components or raster assets defined in Petition 870260008329, dated 01 / 28 / 2026, page 25 / 78 9 / 46 terms of absolute pixel locations. In some cases, a visual asset may include one or more containers, defined in vector space, into which raster assets (e.g., images) can be inserted. A visual asset may include any number of visual components, and a visual asset may be defined in terms of relative positioning or other spatial relationships between visual components within the visual asset. Visual assets may be defined a priori (e.g., as templates or pre-designed components in the 124 graphic design engine) or may be manually designed by a user of the 124 graphic design engine. These spatial relationships between visual components within the visual asset may generally refer to the spacing and / or orientation of one visual component within the visual asset relative to another visual component within the visual asset.

[0029] To accelerate the process of creating visual assets, several modalities described herein utilize generative artificial intelligence models (e.g., generative artificial intelligence model 122) to translate a design objective into a natural language input prompt for modifications applied to an identified visual asset. A generative artificial intelligence model, in general, is an artificial intelligence model that generates content in response to an input prompt specifying parameters based on which the generative artificial intelligence model generates the content. Generative artificial intelligence models can generate content in any of a variety of modalities, such as text content, image content, source code that, when executed, generates a user interface, or other types of content.In general, a design goal can refer to a characteristic (for example, a desired visual appearance or functionality) of a visual asset rendered in a user interface. In many examples, a design goal might specify the desired visual appearance of a visual asset in language used by designers to describe the look and feel of a visual asset and the relationship between components within the visual asset, and / or the desired functionality of one. Petition 870260008329, dated 01 / 28 / 2026, page 26 / 78 10 / 46 A visual resource is defined as a natural language input specifying how the visual resource should react to various inputs or events. For example, a natural language prompt might include design goals indicating that an identified visual resource should be relatively taller or shorter than the current design of the identified visual resource, heavier or lighter than the current design of the visual resource, have more or less space (or breathing room) between visual components than the current design of the visual resource, or similar. In another example, a natural language prompt might include design goals indicating transformations to be applied to a visual resource when various specified events occur, such as changes in the appearance of buttons or other clickable interface elements when such an element is clicked, movement of a visual resource when defined events occur, or similar.

[0030] Furthermore, as discussed in detail in this document, embodiments of the present disclosure may allow the generation of visual assets using forward and backward iteration of designs as an identified base design for future iterations of generating candidate visual asset sets. By allowing forward and backward selection of visual assets as a basis for subsequent iterations of generating candidate visual asset sets, the embodiments presented here may allow generative artificial intelligence models to operate in a non-linear fashion similar to the process by which users design visual assets or generate other content.In other words, unlike generative artificial intelligence models that iterate over the output of the immediately preceding inference round executed by the generative artificial intelligence model, the modalities presented here allow the result of any inference round executed by the generative artificial intelligence model to be used as the basis for which subsequent outputs of the generative artificial intelligence model are generated.

[0031] Figure 2 illustrates an example 200 of a resource Petition 870260008329, dated 01 / 28 / 2026, page 27 / 78 11 / 46 initial visual and an input prompt specifying a design objective for the visual feature, according to some modalities.

[0032] As illustrated, to initialize the generation of candidate visual assets using a generative artificial intelligence model and a natural language input prompt, a visual asset 212 can be added to a user interface 210. The visual asset 212 can be selected from a predefined model or collection of visual assets, or it can be an initial design created by a user of graphic design software. In the example 200 illustrated in Figure 2, the visual asset 212 is illustrated as a horizontal scrolling card that allows a user interface user to view different items presented in a container by scrolling horizontally.

[0033] The natural language input prompt based on which visual feature 212 should be modified can be entered into a control panel 220. As illustrated, the control panel 220 can allow the entry of a natural language prompt specifying a design goal for the selected visual feature 212; as illustrated, the input prompt entered into the control panel 220 specifies that a generative artificial intelligence model should generate candidate visual features that are shorter in height than visual feature 212. In some embodiments, the control panel 220 may further include additional options that can be used by the generative artificial intelligence model to condition how the candidate visual features generated by the generative artificial intelligence model should be generated.For example, as illustrated, control panel 220 might include a slider that can be used to define the range of visual features to be generated based on the generative artificial intelligence model and the natural language input prompt. The slider might indicate, for example, that the set of candidate visual features generated based on the generative artificial intelligence model and the natural language input prompt includes a balanced range of visual features. Petition 870260008329, dated 01 / 28 / 2026, page 28 / 78 12 / 46 candidates or includes a range of candidate visual assets that is biased toward refining a selected visual asset or toward a complete reimagining of the selected visual asset's design.

[0034] Figure 3 illustrates a candidate set of 300 visual features generated by a generative artificial intelligence model based on an input prompt specifying a design objective for a visual feature, according to some modalities.

[0035] The candidate set 300 of visual features can be generated, for example, based on the selected visual feature 212 and an input prompt captured by the control panel 220 illustrated in Figure 2. The input prompt can be a natural language textual prompt entered into a text input field in the control panel 220, an input derived from user interaction with an interactive user interface element in the control panel 220 (e.g., a slider as illustrated in Figure 2), or inputs in any one of a plurality of modalities (voice, image data, eye tracking, motion tracking, etc.). For non-textual inputs, the input prompt can be derived from user interaction with an interactive user interface element, from the extraction of semantic data or an intent from inputs in a non-textual modality, or similar.For example, to generate an input prompt usable by a generative artificial intelligence model to generate the candidate set 300 of visual features, the same or a different generative artificial intelligence model can transform an identified intent of the user interaction with an interactive user interface element or an input in a non-textual input modality into a textual prompt. The candidate set of visual features can include any number of candidate visual features 314, 324, 334 (among others, not illustrated in Figure 3) from which a user can make a selection for use as a basis for further iteration or insertion into a user interface (e.g., the user interface 210 illustrated in Figure 2). In some cases... Petition 870260008329, dated 01 / 28 / 2026, page 29 / 78 13 / 46 modes, the candidate set of 300 visual features may correspond to candidate visual features generated in an initial inference round based on the initial selection of a base visual feature and a natural language prompt processed by the generative artificial intelligence model.

[0036] In general, to generate candidate visual features 314, 324, 334 (among others), the generative artificial intelligence model (e.g., the generative artificial intelligence model 122 illustrated in Figure 1) extracts a design goal from a natural language input prompt and uses the extracted design goal to generate transformations to be applied to the selected visual feature 212. In general, the generative artificial intelligence model can be a design language model trained on a corpus of design-related inputs to translate an identified design goal in a generative artificial intelligence model into a broad set of transformations to be applied to a visual feature. For example, the generative artificial intelligence model can be trained to produce instructions or other outputs that provide a general direction in which visual feature designs should follow.A general direction that visual asset designs should follow might include, for example, a directive output in natural language specifying a relative amount of change to be applied to a selected visual asset 212 based on which one or more transformations can be generated; parameter values ​​specifying a delta between the value of a parameter in the selected visual asset 212 and the value of the corresponding parameter in the candidate visual assets 314, 324, 334; a range of valid adjustments that can be applied to parameters defining the selected visual asset 212 to achieve the specified design goal; or similar. The design language model might identify parameters to be modified based on various concepts expressed in words or collections of words associated with a design goal or direction in an input prompt, such as concepts related to size (smaller, larger, etc.), concepts related to... Petition 870260008329, dated 01 / 28 / 2026, page 30 / 78 14 / 46 spacing (more or less breathing space, more or less visual density, etc.), concepts related to appearance (bolder, lighter, etc.), concepts related to color (hue, luminance, etc.) or similar.

[0037] To visualize candidate visual features 314, 324, 334 (among others not illustrated in Figure 3), the graphic design engine 124 can generate a control panel 310 including visualization options associated with each of the candidate visual features generated by a generative artificial intelligence model based on a design objective extracted from a natural language prompt. The control panel may include buttons or other selectable user interface elements that can allow a user to visualize a candidate visual feature by hovering the cursor over a button associated with the candidate visual feature (for example, to visualize candidate visual feature 314 by hovering the cursor over button 312, to visualize candidate visual feature 324 by hovering the cursor over button 322, to visualize candidate visual feature 334 by hovering the cursor over button 332, etc.).

[0038] Although Figure 3 illustrates candidate visual feature 314 as the same feature as visual feature 212 upon which candidate visual features are generated, it should be understood that candidate visual feature 314 may be a different visual feature generated by modifying visual feature 212. Candidate visual features 324 and 334 illustrate different visual features generated based on different transformations generated from a design objective extracted from a natural language input prompt. In general, candidate visual features 324 and 334 may include the same visual components as visual feature 212, but may display the visual components in a different arrangement according to the design objective extracted from the natural language input prompt.To generate candidate visual feature 324 from visual feature 212 and the natural language input prompt illustrated in Figure 2, one or more transformations generated from the design prompt can specify the container in which. Petition 870260008329, dated 01 / 28 / 2026, page 31 / 78 15 / 46 The visual components of visual resource 212 are arranged lower and can retain the width of visual resource 212. On the other hand, to generate candidate visual resource 334 from visual resource 212 and the natural language input prompt illustrated in Figure 2, one or more transformations generated from the design prompt can alter both the width and height of the container in which the visual components of visual resource 212 are arranged and can alter the order in which the visual components are arranged. For example, due to the change in the width of the container in which the visual components are arranged, candidate visual resource 334 can be generated with a selectable button positioned below a text label, which may be different from the side-by-side arrangement of the text label and the selectable button illustrated in candidate visual resource 324.

[0039] For visual assets that are defined as vector graphics components, the transformations identified by the generative artificial intelligence model may include various changes in the visual properties of these components. These visual properties may include, for example, the size and positioning of one or more visual components associated with the visual asset relative to the size and positioning of these visual components in the selected visual asset 212 based on which the candidate visual assets are generated, changes in textual content, changes in the color palette used to illustrate various components, and so on.For example, to reduce the vertical height of a candidate visual asset relative to the selected visual asset 212, the size of the container and certain components within the container can be reduced by a defined proportional amount (e.g., 50 percent, as illustrated by candidate visual asset 324, 25 percent, as illustrated by candidate visual asset 334, etc.). To adjust the positioning of visual components within a visual asset, the relationship between reference points for different visual components can be adjusted. As illustrated, the relationship between the image component and the other components of the visual asset. Petition 870260008329, dated 01 / 28 / 2026, page 32 / 78 16 / 46 candidate 324 may remain the same, while the relationship between the components in the visual resource candidate 334 may change to reflect the vertical alignment of the image, text label, and selectable button. It should be understood that the above are only examples of transformations that can be identified for application to the selected visual resource 212 to generate the visual resources candidates 314, 324, 334 (among others), and that other transformations (spatial or non-spatial) may be considered.

[0040] Figure 4 illustrates an example 400 of generating different candidate sets of visual features by a generative artificial intelligence model based on the selection of a visual feature from a previous candidate set of visual features, according to some modalities.

[0041] Interface 410 illustrates a control panel in which a set of candidate visual features generated in a first round of inference can be viewed and selected. In interface 410, the candidate visual features may be presented in selection panel 412. In some modalities, selection panel 412 may include a plurality of buttons, with each button corresponding to one of the candidate visual features generated in the first round of inference. As illustrated, a user can select one of the candidate visual features (e.g., the candidate visual feature corresponding to button 414) generated in the first round of inference for use as a basis for further refinement by a generative artificial intelligence model. In some modalities, selecting (e.g., clicking, hovering, etc.)A button corresponding to a candidate visual resource in selection panel 412 can allow a user to preview the appearance of a candidate visual resource in a user interface. The generation of a subsequent set of candidate visual resources can be performed based on selecting a button in selection panel 412 corresponding to a candidate visual resource and executing a command to generate more variants of the selected candidate visual resource. Petition 870260008329, dated 01 / 28 / 2026, page 33 / 78 17 / 46 (for example, clicking a button requesting that the generative artificial intelligence model generate more examples such as the selected candidate visual feature corresponding to button 414).

[0042] Subsequently, a second round of inference is performed to generate a set of candidate visual features presented in selection panel 422. In interface 420, the first set of candidate visual features presented in selection panel 412 can be preserved so that a user can change the selection of visual features included in a user interface to a visual feature different from the visual feature corresponding to button 414. As illustrated, selection panel 422, like selection panel 412 discussed earlier, can include a plurality of buttons that are associated with the set of candidate visual features generated during the second round of inference. Selecting a button (e.g., button 424) can be used to preview the appearance of a candidate visual feature in the user interface.As discussed earlier with regard to interface 410, the generation of a subsequent set of candidate visual assets can be performed based on a button selection in the selection panel 422 and the execution of a command to generate more variants of the selected candidate visual asset.

[0043] Interface 430 illustrates a control panel in which candidate visuals generated in a third round of inference can be viewed and selected. The candidate visuals generated in the third round of inference, which can be displayed in selection panel 432, can be generated based on the selection of a visual corresponding to button 424 in the second round of inference (and indirectly from the selection of the visual corresponding to button 414 in the first round of inference). As illustrated, interface 430 can include selection panels 412 and 422 to allow a user to select a visual from a previous round of inference as a basis for generating subsequent variants in subsequent rounds of inference using the intelligence model. Petition 870260008329, dated 01 / 28 / 2026, page 34 / 78 18 / 46 artificial generative.

[0044] In some embodiments, the amount by which candidate visual features are generated in an inference round using a generative model (e.g., the generative model 122 illustrated in Figure 1) can change based on the number of previous inference rounds performed to design a specific visual feature. For example, in the first inference round, a wide variety of candidate visual features may be generated, with each candidate visual feature having significantly different appearances (e.g., as illustrated in Figure 3, where candidate visual features 314, 324, 334 have significantly different height and / or width parameters and different arrangements of visual components). Subsequent inference rounds may be performed to generate refinements to the selected visual feature generated in a previous inference round.For example, the candidate visuals included in selection panel 422 may include variations generated from the visual corresponding to button 414 using smaller transformations than those used to generate the candidate visuals included in selection panel 412. Meanwhile, the candidate visuals included in selection panel 432 may include variations generated from the visual corresponding to button 424 using smaller transformations than those used to generate the candidate visuals included in selection panel 422.

[0045] In some embodiments, a generative model (for example, the generative model 122 illustrated in Figure 1) can further refine a selected candidate visual feature based on another natural language prompt embedded in the generative artificial intelligence model. The natural language prompt used to refine the selected candidate visual feature may include additional design objective information based on which refinements can be generated. In some embodiments, the design objective information specified in the natural language prompt Petition 870260008329, dated 01 / 28 / 2026, page 35 / 78 19 / 46 natural language prompts can reference another visual feature in the user interface. When the natural language prompt references another visual feature in the user interface, the generative artificial intelligence model can ingest parameters associated with the referenced visual feature and determine transformations to be used in generating candidate visual features based on the parameters associated with the referenced visual feature. For example, a natural language prompt might specify that the design of a visual feature should be less visually prominent than a referenced visual feature. To be less visually prominent, several transformations related to the size and weight of the visual components in the visual feature can be generated.In general, these transformations can reduce the size of various visual components in the visual asset to a size smaller than the referenced visual asset, reduce the visual weight of the text displayed in the visual asset (for example, changing from a bolder version of a typeface to a lighter version of the typeface), and so on.

[0046] In some embodiments, the transformations generated by the generative artificial intelligence model (e.g., the generative artificial intelligence model 122 illustrated in Figure 1) to generate candidate visual features may be constrained by several parameters. For example, visual components including textual data may have a minimum height that can be defined a priori based on, for example, the resolution of a screen on which a user interface is to be rendered, the physical size of the screen, the pixel density of the screen, and the like. Transformations applied to these visual components may perform unlimited modifications to a variety of parameters associated with these visual components that do not affect the height of a visual component (e.g., they may perform any of a variety of transformations related to positioning, font weight, typeface, etc.).Transformations related to height applied to these visual components, however, may be restricted to... Petition 870260008329, dated 01 / 28 / 2026, page 36 / 78 20 / 46 that the height of these visual components in a candidate visual asset does not fall below the minimum height defined a priori.

[0047] In some embodiments, user-directed modifications to a visual feature can be performed before generating additional candidate variants of the visual feature using transformations generated based on a design objective extracted from a natural language prompt by a generative artificial intelligence model. For example, a user can add, remove, or modify the appearance of various visual components in a visual feature at any time during the process of generating a visual feature using natural language prompts and a generative artificial intelligence model. The modifications to a candidate visual feature generated in the i-th round of inference can subsequently be used as the basis upon which candidate visual features are generated in the i+1st round of inference.

[0048] Figure 5 is a flowchart illustrating exemplary operations 500 for generating visual features based on an input prompt specifying a design objective for a visual feature and selection of the visual feature based on which a generative artificial intelligence model (e.g., the generative model 122 illustrated in Figure 1) generates the visual features, according to some modalities. The operations 500 can be performed, for example, by a computational system including one or more processors, such as the processor(s) 102 of the computational device 100 illustrated in Figure 1.

[0049] As illustrated, operations 500 can begin in block 510, in which a processor receives an indication of a visual feature for modification and a natural language input prompt specifying a design goal to be applied to the indicated visual feature. The natural language input prompt can be provided directly to the processor or can be derived by the processor based on other inputs to the processor. For example, a natural language input prompt Petition 870260008329, dated 01 / 28 / 2026, page 37 / 78 21 / 46 natural can be derived based on an intent or other semantic information extracted from user interaction with interactive user interface elements (e.g., sliders, buttons, etc. associated with changes in specific properties of a visual feature), inputs in non-textual modalities (e.g., audio and / or visual inputs), or similar.

[0050] In block 520, operations 500 continue with the processor translating the design objective into a plurality of transformations to be applied to the indicated visual resource.

[0051] In some modalities, the plurality of transformations comprises modifications in a vector definition of the indicated visual resource that result in the creation of visual resources in accordance with the design objective. In some modalities, the modifications comprise one or more of a change in size, a change in position, or a change in spatial relationships with respect to one or more other visual components in the visual resource.

[0052] In block 530, operations 500 proceed with the generation of a candidate set of visual resources based on the application of plurality of transformations to the indicated visual resource.

[0053] In some embodiments, the candidate set of visual features comprises a first set of visual features generated based on the application of plurality of transformations to the indicated visual feature and one or more other sets of visual features generated based on previous iterations of applying transformations to previous selections of visual features. As discussed earlier, the magnitude of the transformations used in different iterations (e.g., inference rounds) may differ based on the selection of a change amount to be applied to a selected visual feature generated in a previous inference round and the number of inference rounds previously executed. Larger transformations may be applied in rounds of Petition 870260008329, dated 01 / 28 / 2026, page 38 / 78 22 / 46 previous inference rounds or when a user requests a significant reimagining of a visual asset, while minor transformations may be applied in later inference rounds or when a user specifies that the next inference round should generate minor refinements from a selected visual asset.

[0054] In block 540, operations 500 proceed with the output of the candidate set of visual features for selection in a user interface.

[0055] In some embodiments, the indication of the visual feature comprises a selection of the visual feature from a candidate set of visual features, the candidate set including one or more subsets of visual features generated based on transformations generated from the design objective. Visual features in a second subset of visual features in the one or more subsets may be generated based on a visual feature selected in a first subset of visual features.

[0056] In some embodiments, the indicated visual feature comprises a visual feature included in a first subset of visual features, the first subset of visual features being generated in an inference round prior to an inference round in which a second subset of visual features was generated. 500 operations may further include the generation of the candidate set of visual features based on the generation of a third subset of visual features based on the indicated visual feature in the first subset of visual features and the plurality of transformations. To produce the output of the candidate set of visual features for selection in the user interface, the second subset of visual features may be replaced by the third subset of visual features.

[0057] In some embodiments, 500 operations also include temporarily modifying the user interface based on a transient selection of a visual feature in the candidate set of visual features. For example, when producing the output of the candidate set of features. Petition 870260008329, dated 01 / 28 / 2026, page 39 / 78 23 / 46 visuals for selection in a user interface, each visual resource in the candidate set of visual resources can be rendered in thumbnail form in a selectable user interface control (e.g., a button, radio button, dropdown menu, etc.). Hovering the cursor over a user interface control associated with a candidate visual resource can cause the candidate visual resource to be rendered in a user interface and displayed until a user moves the cursor away from the user interface control associated with the candidate visual resource.

[0058] In some embodiments, the 500 operations also include receiving an indication of a selected visual resource from the candidate set of visual resources to implement in the user interface. An indication of one or more modifications to be applied to the selected visual resource may be received, and a modified visual resource may be generated based on the one or more indicated modifications and the selected visual resource. In some embodiments, the one or more indicated modifications comprise replacing placeholder content in the selected visual resource with other content. In some embodiments, the one or more indicated modifications comprise additional visual components to be added to the selected visual resource, the additional visual components being defined based on at least one position in which the additional visual components should be added to the selected visual resource and a size of the additional visual components. Example of Visual Resource Modification Based on Generative Artificial Intelligence Models

[0059] Design objectives extracted from a natural language input prompt, as discussed, can be used to generate transformations to be applied to visual components in a visual asset. Because these transformations are applied to specific properties of different visual components in a visual asset, the generation of transformations to be applied to visual components in a visual asset Petition 870260008329, dated 01 / 28 / 2026, page 40 / 78 24 / 46 can be used to populate one or more control panels that allow user-directed modification of a visual feature.

[0060] Figure 6 illustrates control panels for adjusting the appearance of a visual resource based on an input prompt specifying a design objective for the visual resource, according to some modalities.

[0061] As illustrated, a visual feature 610, which may have been generated from a natural language input prompt as discussed earlier or may be a predefined visual feature, can be displayed in a user interface. A natural language prompt (not illustrated in Figure 6) can be entered into a generative artificial intelligence model to specify a design goal for the visual feature 610. In example 600 illustrated in Figure 6, the natural language prompt can specify a density and visual weight goal to be modified. Generally, modifications to density design goals may include modifications to the spacing between various visual components of a visual feature. Modifications to visual weight design goals may include modifications to text size and / or style, line weight and / or style, and the like.

[0062] To allow modification of visual feature 610, a graphic design engine (e.g., graphic design engine 124 illustrated in Figure 1) can populate one or more control panels with various controls based on the design objective extracted from a natural language input prompt by a generative model (e.g., generative model 122 illustrated in Figure 1). In control panel 620, sliders 622 and 624 can be used to apply a set of transformations to visual feature 610 and / or set minimum and maximum values ​​for parameters in control panel 630. Control panel 630, in general, allows user-defined modifications to parameters of visual components associated with the visual feature based on manual input or selection of Petition 870260008329, dated 01 / 28 / 2026, page 41 / 78 25 / 46 values ​​for these parameters.

[0063] When a visual asset is selected, the values ​​for the parameters associated with the visual asset components that are relevant to the design objective specified in the natural language prompt may be populated by the graphic design engine in the relevant portions of control panel 630. Control panel 630 may include parameters relevant to the visual components of the visual asset. In example 600, the visual components for which parameters are displayed in control panel 630 include the container in which other visual components of the visual asset are displayed and the other visual components of the visual asset. The container may correspond to the card component illustrated in control panel 630, and the other visual components of the visual asset contained within the card component may include the description, thumbnail image, labels, and buttons.

[0064] Figure 7 illustrates an example 700 of changes to a visual resource generated based on adjustments to one or more parameters in control panels to adjust the appearance of a visual resource, according to some modalities.

[0065] Visual feature 710 can be generated by the graphic design engine (e.g., graphic design engine 124 illustrated in Figure 1) by applying changes to visual feature 610 illustrated in Figure 6 by changing the position of slider 622 to a new position illustrated by slider 722 in control panel 720. As illustrated, visual feature 710 can be a less dense version of visual feature 610. When the value of a slider changes, a transformation can be identified by a generative model (e.g., generative model 122 illustrated in Figure 1) that achieves a desired design goal. For example, to generate a visual feature that is less dense, the identified transformation might increase the size of a container associated with a visual feature and might increase the spacing between different Petition 870260008329, dated 01 / 28 / 2026, page 42 / 78 26 / 46 visual components included in the container. In some embodiments, when the position of a slider changes from that illustrated by slider 622 in Figure 6 to that illustrated by slider 722 in Figure 7, the graphic design engine can apply visual effects in control panel 730 to the parameter values ​​associated with the visual components to allow a user to easily identify which parameters have been changed.

[0066] Figure 8 illustrates an example 800 showing changes in a visual resource generated based on adjustments to parameters identified based on an input prompt specifying a design objective for the visual resource, according to some modalities.

[0067] In example 800, a base visual feature 810 can be modified based on input from a natural language prompt in a generative artificial intelligence model (e.g., the generative model 122 illustrated in Figure 1) via a control panel 812. As discussed with respect to the control panel 220 illustrated in Figure 2, a natural language prompt specifying a design objective and a range of variants to be generated by a generative artificial intelligence model can be specified. As illustrated, the natural language prompt specifies that the base visual feature 810 should be modified to be made more fun and that different cards should be rotated based on the depth of a card in a stack of cards.

[0068] A candidate visual asset 820 can be generated by the graphic design engine based on transformations identified from the base visual asset 810 and the design objective extracted from the natural language input prompt. As described above with respect to Figures 6 and 7, the properties relevant to the design objective extracted from the natural language input prompt can be populated in a control panel 830. In example 800, the properties identified from the design objective might include the depth distance between cards, information Petition 870260008329, dated 01 / 28 / 2026, page 43 / 78 27 / 46 rotations for different cards, a range of valid rotation values, elevation data, and the like. A user can accept the properties generated from the transformation generated by the generative artificial intelligence model or apply various changes to the properties of the visual components by modifying the values ​​associated with these properties in the 830 control panel. Modifications to these values ​​can be performed based on direct input of new values ​​in text boxes associated with different properties of the visual components, verbal and / or image-based inputs in a generative artificial intelligence model, input from a text prompt in a generative artificial intelligence model, etc. In some embodiments, the properties of the visual components can be associated with various user interface controls (e.g., sliders, incrementers, etc.).), the interaction with which can modify the values ​​associated with the properties of the visual components.

[0069] Figure 9 is a flowchart illustrating exemplary operations 900 for modifying visual features based on changes in parameters identified by a generative artificial intelligence model from an input prompt specifying a design objective for a visual feature, according to some modalities. The operations 900 can be performed, for example, by a computational system including one or more processors, such as the processor(s) 102 of the computational device 100 illustrated in Figure 1.

[0070] As illustrated, operations 900 can begin in block 910, where a processor receives an indication of a visual feature for modification and a natural language input prompt specifying a design goal to be applied to the indicated visual feature.

[0071] In some embodiments, the input comprises a slider input associated with a design property of the visual feature. The slider input may include a plurality of positions, with each respective position of the plurality of positions being Petition 870260008329, dated 01 / 28 / 2026, page 44 / 78 28 / 46 associated with a design objective to be applied to the indicated visual asset. For example, a slider that adjusts the density of a visual asset might include several positions corresponding to more or less density for the visual asset. In another example, a slider that adjusts the visual weight of a visual asset might include several positions corresponding to heavier or lighter weight for text, borders, and similar elements for the visual asset.

[0072] In some modes, an adjustment to the slider input applies a global modification to the identified visual components. The global modification can modify properties associated with multiple visual components associated with a visual resource.

[0073] In some modes, the slider input is displayed in a first control panel. An adjustment to the slider input modifies one or more of the properties of the identified visual components displayed in a second control panel.

[0074] In some embodiments, a range associated with the slider input may include a minimum value associated with a first set of values ​​for the identified properties of the identified visual components and a maximum value associated with a second set of values ​​for the identified properties of the identified visual components. The first set of values ​​and the second set of values ​​may be identified based on the specified design objective.

[0075] In some modalities, the input comprises an adjustment to a property of the identified properties associated with a visual component of the identified visual components.

[0076] In some modalities, a range of valid values ​​for the identified properties of the identified visual components is associated with a defined amount of modification to be applied to the visual resource.

[0077] In some modalities, a range of valid values ​​for the identified properties of the visual components Petition 870260008329, dated 01 / 28 / 2026, page 45 / 78 29 / 46 identified is defined in relation to another visual component in a user interface.

[0078] In block 920, operations 900 proceed with the identification of visual components in the indicated visual resource and properties of the identified visual components to be modified based on the specified design objective.

[0079] In block 930, operations 900 proceed with populating one or more control panels with one or more controls to modify the identified visual components to achieve the specified design objective.

[0080] In block 940, operations 900 proceed with the receipt of an entry in at least one of the one or more control panels.

[0081] In block 950, operations 900 proceed with the modification of the indicated visual resource based on the received input.

[0082] By using generative artificial intelligence models to identify design objectives in a natural language prompt and modify visual assets based on the identified design objectives, the modalities presented here can allow generative artificial intelligence models to accelerate the process of creating visual assets. Generative artificial intelligence models can be used to generate a variety of candidate visual assets based on which additional sets of candidate visual assets can be generated. Furthermore, a history of candidate visual assets generated by the generative artificial intelligence model can be maintained so that a user can move forward and backward between generated visual assets during a variety of inference rounds performed by the generative artificial intelligence model.By doing this, the modalities presented here can allow generative artificial intelligence models to perform inference in a non-linear way in which the output of any previous inference round can be used as context. Petition 870260008329, dated 01 / 28 / 2026, page 46 / 78 30 / 46 for the execution of a new round of inference. Furthermore, identifying a design objective and generating transformations to be applied to a visual feature can allow visual features to be designed using both coarse and refined modifications to parameters associated with the visual feature.

[0083] Figure 10 illustrates a networked computing system for implementing an interactive application platform on a user computing device, according to some embodiments. A networked computing system as illustrated in Figure 10 can be implemented using one or more servers that communicate with user computing devices through one or more networks. The networked computing system 1050 illustrated in Figure 10 may correspond, for example, to the computing device 100 illustrated in Figure 1 and may be used to generate and / or modify visual content based on generative artificial intelligence models and an input prompt specifying a design objective for the visual content.

[0084] In some embodiments, the networked computing system 1050 performs operations to enable an interactive application platform (IAP 1000) to be implemented on user computing devices 10. In some embodiments, the IAP 1000 can be implemented by the user initiating a session (e.g., user accessing a website) to receive programmatic resources from the IAP 1000. A browser component executes the programmatic resources to implement the IAP 1000, with functionality to receive user input and render content that is based on or responsive to user input. As described, the IAP 1000 is implemented to enable the user to create various types of content, such as interactive graphic designs, art, whiteboard content, program code renderings, presentations, and / or textual content.As described further, IAP 1000 can include logic (ASL 1016) to implement one or more application services, where each service... Petition 870260008329, dated 01 / 28 / 2026, page 47 / 78 31 / 46 The application is implemented through IAP 1000 to provide a matching set of functionality and user experience. IAP 1000 also implements application services to share some resources, such as the screen, workspace file, or design element library. Furthermore, IAP 1000 allows multiple application services to be used during a given online session and / or in relation to a specific application service.

[0085] According to some embodiments, a user of the computing device 10 operates the web-based application 80 to access a network site, where programmatic resources are retrieved and executed to implement the IAP 1000. The web-based application 80 may execute scripts, code, and / or other logic (the programmatic components) to implement the functionality of the IAP 1000. In some embodiments, the web-based application 80 may correspond to a commercially available browser, such as GOOGLE CHROME (developed by GOOGLE, INC.) or SAFARI (developed by APPLE, INC.). In some embodiments, the IAP 1000 processes may be implemented as scripts and / or other embedded code that the web-based application 80 downloads from a network site. For example, the web-based application 80 may execute code that is embedded in a web page to implement IAP 1000 processes.The web-based application 80 can also execute scripts to retrieve other scripts and programmatic resources (e.g., libraries) from the web site and / or other local or remote locations. For example, the web-based application 80 can execute JavaScript embedded in an HTML resource (e.g., a web page structured according to HTML 5.0 or other versions, as provided under standards published by the W3C or WHATWG consortia). In some embodiments, the rendering engine 1020 can utilize graphics processing unit (GPU) accelerated logic, as provided through WebGL (Web Graphics Library) programs that execute GLSL (Graphics Library Shader Language) programs. Petition 870260008329, dated 01 / 28 / 2026, page 48 / 78 32 / 46 that run on GPUs.

[0086] IAP 1000 can be implemented as part of a network service, where the web-based application 80 communicates with one or more remote computers (e.g., a server used for a network service) to execute IAP 1000 processes. The web-based application 80 retrieves some or all of the programmatic resources to implement IAP 1000 from a network site. The web-based application 80 can also access various types of datasets when providing IAP 1000. Datasets can correspond to files and design libraries (e.g., pre-designed design elements), which can be stored remotely (e.g., on a server, in association with an account) or locally. In some embodiments, the networked computing system 1050 provides a shared design library that the user computing device 10 can utilize with any of the application services provided through IAP 1000.In this way, the user can start a session to implement IAP 1000 for the purpose of creating or editing the workspace file, as rendered on screen 1022, according to one of the multiple collaborative application services of IAP 1000.

[0087] In some embodiments, IAP 1000 includes a program interface 1002, an input interface 1018, and a rendering engine 1020. The program interface 1002 may include one or more processes that are executed to access and retrieve programmatic resources from local and / or remote sources. In one implementation, the program interface 1002 may, for example, generate a screen 1022, using programmatic resources that are associated with the web-based application 80 (e.g., HTML 5.0 screen). As an addition or variation, the program interface 1002 may trigger or otherwise cause the screen 1022 to be generated using programmatic resources and datasets (e.g., screen parameters) that are retrieved from local (e.g., memory) or remote (e.g., network service) sources. Petition 870260008329, dated 01 / 28 / 2026, page 49 / 78 33 / 46

[0088] Program interface 1002 may also retrieve programmatic resources that include an application structure for use with screen 1022. The application structure may include datasets that define or configure, for example, a set of interactive graphical tools that integrate with screen 1022 and that comprise input interface 1018, to allow the user to provide input to generate or update content rendered on screen 1022.

[0089] According to some embodiments, the input interface 1018 can be implemented as a functional layer that is integrated with the screen 1022 to detect and interpret user input. The input interface 1018 can, for example, process a user's interaction with an input mechanism (e.g., pointing device, keyboard) of the user's computing device, to detect, for example, cursor positioning / movement relative to the screen 1022, cursor hover input (e.g., pre-selection input), selection input (e.g., clicks or double-clicks), shortcuts (e.g., keyboard inputs), and other inputs. When processing a user's interaction with a pointing device, the input interface 1018 can use a reference from the screen 1022 to identify a screen location of a user's cursor as the user moves or otherwise interacts with the pointing device.Additionally, the 1018 input interface can interpret a user input action based on the location of the detected input (e.g., whether the input position indicates selection of a tool, a rendered object on the screen, or a region of the screen), the frequency of the detected input within a given time period (e.g., double-click), and / or the start and end positions of an input or series of inputs (e.g., start and end positions of a click and drag), as well as various other types of input that the user can specify (e.g., right-click, screen touch, etc.) through one or more input devices. In some embodiments, the 1018 input interface can interpret, for example, a series of inputs. Petition 870260008329, dated 01 / 28 / 2026, page 50 / 78 34 / 46 as a design tool selection (e.g., shape selection based on input location), as well as inputs to define attributes (e.g., dimensions) of a selected shape. In some embodiments, the 1018 input interface may interpret a continuous input (corresponding to a continuous movement of the user's pointing device) as selecting a tool (e.g., shape tool) and screen location where an output of the selected tool should appear.

[0090] In some embodiments, IAP 1000 includes application service logic 1016 to allow multiple application services to be used during a given user session, where each application service provides the user with a particular functionality and / or user experience. As described by some embodiments, each application service is implemented by IAP 1000 using a corresponding application service logic 1016 to configure the interface component 1018, rendering engine 1020, and / or other IAP 1000 components to provide the functionality and user experience of the corresponding application service. In this way, IAP 1000 allows the user to operate multiple application services during an online session. Furthermore, the different application services can share resources, including programmatic resources of IAP 1000, such as the screen 1022.In this way, each application service can contribute content to and / or utilize resources and content provided with screen 1022 during a given session. Furthermore, application services can be implemented as alternative modes of IAP 1000, so that the user can switch between modes, where each mode provides a particular functionality and user experience. In some modes, each application service can utilize a common workspace file associated with the user. By default, a computing device that opens the workspace file can use a standard application service to access and / or update that workspace file. The user can also switch the IAP 1000 operating mode to utilize one. Petition 870260008329, dated 01 / 28 / 2026, page 51 / 78 35 / 46 different application service to access, use and / or update the workspace file.

[0091] The networked computing system 1050 may include a site manager 1058 for managing a website where a set of web resources 1055 (e.g., web page) are made available to the web-based application 80 of the user computing devices 10. The web resources 1055 may include instructions, such as scripts or other logic (ICAP instructions 1057), that are executable by browsers or web components of user computing devices. The web resources 1055 may also include (i) resources that must be shared between application services, provided to user computing devices in connection with user computing devices using any of the application services, and (ii) application-specific resources, which run on user computing devices for a particular of the available application services.Web resources 1055 may also include a design library of design elements that is partially or fully shared among application services. The design element library may allow the user to select predefined design elements for use on the screen 1022 in connection with the user using any of the application services.

[0092] In some variations, once the computing device 10 accesses and downloads web resources 1055, the web-based application 80 executes IAP instructions 1057 to implement the functionality as described previously. For example, IAP instructions 1057 may be executed by the web-based application 80 to start program interface 1002 on the user computing device 10. The initiation of program interface 1002 may coincide with the establishment of, for example, a web-socket connection between program interface 1002 and a service component 1060 of the networked computing system 1050.

[0093] In some modes, web resources 1055 Petition 870260008329, dated 01 / 28 / 2026, page 52 / 78 36 / 46 include logic that the web-based application 80 executes to initiate one or more processes of a program interface 1002, causing the IAP 1000 to retrieve additional programmatic resources and datasets to implement the functionality as described by the examples. The web resources 1055 may, for example, incorporate logic (e.g., JAVASCRIPT code), including GPU-accelerated logic, into an HTML page for download by users' computing devices. The program interface 1002 may be triggered to retrieve additional programmatic resources and datasets from, for example, the network service 1052, and / or local resources of the computing device 10, in order to implement each of the multiple application services of the IAP 1000.For example, some of the IAP 1000 components can be implemented through web pages that can be downloaded to computing device 10 after authentication is performed, and / or once the user performs additional actions (e.g., downloading one or more pages from the workspace associated with the account identifier). Consequently, in the examples as described, networked computing system 1050 can communicate IAP 1057 instructions to computing device 10 through a combination of network communications, including through download activity of the web-based application 80, where IAP 1057 instructions are received and executed by the web-based application 80.

[0094] The computing device 10 can use the web-based application 80 to access a network service site 1052 to download the web page or web resource. When accessing the site, the web-based application 80 can automatically (e.g., through saved credentials) or through manual input, communicate an account identifier to the service component 1060. In some embodiments, the web-based application 80 can also communicate one or more additional identifiers that correlate to a user identifier.

[0095] Additionally, in some modalities, the Petition 870260008329, dated 01 / 28 / 2026, p. 53 / 78 37 / 46 service component 1060 can use the user identifier or user identifier account to retrieve profile information 1009 from a user profile store 1066. As an addition or variation, profile information 1009 for the user can be determined and stored locally on the user's computing device 10.

[0096] The 1060 service component can also retrieve files from an active workspace (1063 active workspace files) that are linked to the account or user identifier of a 1064 file store. The 1066 profile store can also identify the workspace that is identified with the account and / or user, and the 1064 file store can store the datasets that comprise the workspace. The datasets stored with the 1064 file store may include, for example, the pages of a workspace and one or more 1061 data structure representations for the design being edited that is renderable from the respective active workspace files.

[0097] As an addition or variation, each file can be associated with metadata that identifies the application service that was used to create the particular file. In some embodiments, the metadata identifies the default application service for viewing, using, or otherwise updating the application service.

[0098] Additionally, in some embodiments, the service component 1060 provides a representation 1059 of the user-associated workspace for the web-based application 80, where the representation identifies, for example, individual files associated with the user and / or user account. The workspace representation 1059 may also identify a set of files, where each file includes one or more pages, and each page includes objects that are part of a design interface.

[0099] On user device 10, the user can Petition 870260008329, dated 01 / 28 / 2026, page 54 / 78 38 / 46 view the workspace representation through the web-based application 80, and the user can choose to open a workspace file through the web-based application 80. In some embodiments, when the user chooses to open one of the active workspace files 1063, the web-based application 80 starts the screen 1022. For example, the IAP 1000 can start an HTML 5.0 screen as a component of the web-based application 80, and the rendering engine 1020 can access one or more representations of data structures 1011 of content rendered on the screen 1022.

[00100] The IAP 1000 uses application service logic 1016 to implement multiple operating modes, where each mode corresponds to an application service. As described, the application service logic 1016 associated with each service application can include instructions and data to configure the IAP 1000 components to include functionality and features of the corresponding application service. Consequently, the application service logic 1016 can, for example, configure the application structure and / or input interface 1018 to differ in form, functionality, and / or configuration between the alternative modes of the IAP 1000. Additionally, the type of actions and interactions that the user can perform to register input can vary based on the operating modes. Furthermore, the different operating modes can include different input features or user interface for the user to select and use for screen inclusion 1022.For example, when the IAP 1000 is operating in a whiteboard service application mode, the program interface 1002 may provide input features to allow a user to select a design element that is in the form of a sticky note, whereas, in an alternative mode for an interactive graphic design service application, the sticky note feature is not available. However, in the alternative mode, the user may be able to select any of numerous possible shapes or pre-designed objects on which the user can write text messages for display. Petition 870260008329, dated 01 / 28 / 2026, page 55 / 78 39 / 46 on screen 1022.

[00101] Additionally, application service logic 1016 can configure the operation of rendering engine 1020, so that the functionality and behavior of rendering engine 1020 differ between different application services. In this way, rendering engine 1020 works to provide alternative behaviors for different operating modes, coinciding with the particular service application that is active.For example, the rendering engine configuration 1020 can affect the appearance of the screen 1022, the appearance (e.g., visual attributes) of content elements rendered on the screen 1022, the behavior or representation of user interaction (e.g., whether the user's cursor or pointing device is represented on the screen 1022), the specific type or content that is rendered, the physics engine used by the rendering engine to represent dynamic events (e.g., the object being moved), what user operations can be performed (e.g., whether selected objects can be resized), and the like.

[00102] Additionally, each of the application services can utilize a shared library of content elements (e.g., graphic design element), as well as core functionality that allows design elements to be shared and updated between the different application services available through the platform. Furthermore, a workspace file created and edited using one application service can be used with another application service. Moreover, the transition between application services can be seamless—for example, user 10's computing device can open a workspace file using the first application service (e.g., interactive graphic design application service for UIX design) and then seamlessly switch to using the second application service (e.g., whiteboard application service) with the same file, without closing the file. Petition 870260008329, dated 01 / 28 / 2026, page 56 / 78 40 / 46 workspace. In some modes, each application service allows the user to update the workspace file even while the workspace file is in use by other computing devices (for example, as in a collaborative environment). In some modes, the user can switch the IAP 1000 mode to change which application service is in use, with each application service utilizing the workspace file. Exemplary Clauses

[00103] Several aspects of this disclosure are described in the following numbered clauses.

[00104] 1. In some embodiments, a processor-implemented method comprising: receiving an indication of a visual feature for modification and a natural language input prompt specifying a design objective to be applied to the indicated visual feature; translating the design objective into a plurality of transformations to be applied to the indicated visual feature; generating a candidate set of visual features based on the application of the plurality of transformations to the indicated visual feature; and producing the output of the candidate set of visual features for selection in a user interface.

[00105] 2: The method of clause 1, wherein the candidate set of visual features comprises a first set of visual features generated based on the application of plurality of transformations to the indicated visual feature and one or more other sets of visual features generated based on previous iterations of applying transformations to previous selections of visual features.

[00106] 3: The method of clauses 1 or 2, where the indication of the visual resource comprises a selection of the visual resource from a candidate set of visual resources, the candidate set including one or more subsets of visual resources generated based on transformations generated from the design objective. Petition 870260008329, dated 01 / 28 / 2026, page 57 / 78 41 / 46

[00107] 4: The method of clause 3, in which visual assets in a second subset of visual assets in one or more subsets are generated based on a visual asset selected in a first subset of visual assets.

[00108] 5: The method of clauses 3 or 4, wherein: the indicated visual feature comprises a visual feature included in a first subset of visual features, the first subset of visual features being generated in an inference round prior to an inference round in which a second subset of visual features was generated; generating the candidate set of visual features comprises generating a third subset of visual features based on the indicated visual feature in the first subset of visual features and the plurality of transformations; and producing the output of the candidate set of visual features for selection in the user interface comprises replacing the second subset of visual features with the third subset of visual features.

[00109] 6: The method of any of clauses 1 to 5, further comprising: temporarily modifying the user interface based on a transient selection of a visual feature from the candidate set of visual features.

[00110] 7: The method of clauses 1 to 6, further comprising: receiving an indication of a selected visual resource from the candidate set of visual resources to implement in the user interface; receiving an indication of one or more modifications to be applied to the selected visual resource; and generating a modified visual resource based on the one or more indicated modifications and the selected visual resource.

[00111] 8: The method of clause 7, in which one or more of the indicated modifications comprise the replacement of placeholder content in the selected visual resource with other content.

[00112] 9: The method of clauses 7 or 8, in which one or more of the indicated modifications comprise additional visual components to Petition 870260008329, dated 01 / 28 / 2026, page 58 / 78 42 / 46 to be added to the selected visual asset, the additional visual components being defined based on at least one position where the additional visual components should be added to the selected visual asset and a size of the additional visual components.

[00113] 10: The method of any of clauses 1 to 9, in which the plurality of transformations comprises modifications in a vector definition of the indicated visual resource that result in the creation of visual resources in accordance with the design objective.

[00114] 11: The method of clause 10, in which the modifications comprise one or more of a change in size, a change in position, or a change in spatial relationships with respect to one or more other visual components in the visual resource.

[00115] 12: A processor-implemented method comprising: receiving an indication of a visual resource for modification and a natural language input prompt specifying a design objective to be applied to the indicated visual resource; identifying visual components in the indicated visual resource and properties of the identified visual components to be modified based on the specified design objective; populating one or more control panels with one or more controls to modify the identified visual components to achieve the specified design objective; receiving input in at least one of the one or more control panels; and modifying the indicated visual resource based on the received input.

[00116] 13: The method of clause 12, in which the input comprises a slider input associated with a design property of the visual resource.

[00117] 14: The method of clause 13, in which an adjustment to the slider input applies a global modification to the identified visual components.

[00118] 15: The method of clauses 13 or 14, where the slider input is displayed in a first control panel, and in Petition 870260008329, dated 01 / 28 / 2026, page 59 / 78 43 / 46 that an adjustment in the slider input modifies one or more of the properties of the identified visual components displayed in a second control panel.

[00119] 16: The method of any of clauses 13 to 15, wherein a range associated with the slider input comprises a minimum value associated with a first set of values ​​for the identified properties of the identified visual components and a maximum value associated with a second set of values ​​for the identified properties of the identified visual components, and wherein the first set of values ​​and the second set of values ​​are identified based on the specified design objective.

[00120] 17: The method of any of clauses 12 to 16, where the input comprises an adjustment to a property of the identified properties associated with a visual component of the identified visual components.

[00121] 18: The method of any of clauses 12 to 17, in which a range of valid values ​​for the identified properties of the identified visual components is associated with a defined amount of modification to be applied to the visual resource.

[00122] 19: The method of any of clauses 12 to 18, in which a range of valid values ​​for the identified properties of the identified visual components is defined in relation to another visual component in a user interface.

[00123] 20. A processing system comprising: at least one memory having executable instructions in it; and one or more processors configured to execute the executable instructions to cause the processing system to perform the method of any of clauses 1 to 19.

[00124] 21. A non-transient, computer-readable medium having executable instructions stored thereon that, when Petition 870260008329, dated 01 / 28 / 2026, pp. 60 / 78 44 / 46 processed by one or more processors, cause one or more processors to perform the method of any of the clauses 1 to 19.

[00125] Any and all combinations of any of the claim elements cited in any of the claims and / or any elements described in this application, in any form, are within the contemplated scope of the present invention and protection.

[00126] The descriptions of the various modalities have been presented for illustrative purposes, but are not intended to be exhaustive or limited to the modalities disclosed. Many modifications and variations will become apparent to those skilled in the art without departing from the scope of the modalities described.

[00127] Aspects of the present embodiments may be incorporated as a computer program system, method, or product. Consequently, aspects of this disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, microcode, etc.), or an embodiment combining software and hardware aspects, all of which may generally be referred to herein as a module, a system, or a computer. Furthermore, any technique, process, function, component, mechanism, module, or hardware and / or software system described in this disclosure may be implemented as a circuit or set of circuits. In addition, aspects of this disclosure may take the form of a computer program product embodied in one or more computer-readable media having computer-readable program code embedded therein.

[00128] Any combination of one or more computer-readable media may be used. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example, but not limited to, an electronic system, apparatus or device, Petition 870260008329, dated 01 / 28 / 2026, pp. 61 / 78 45 / 46 magnetic, optical, electromagnetic, infrared, or semiconductor, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media would include the following: an electrical connection having one or more wires, a portable computer floppy disk, a hard disk drive, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc (CD-ROM) read-only memory, an optical storage device, a magnetic storage device, or any suitable combination thereof. In the context of this document, a computer-readable storage medium may be any tangible medium that can contain or store a program for use by or in connection with an instruction-executing system, apparatus, or device.

[00129] Aspects of the present disclosure are described above with reference to flowchart illustrations and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be supplied to a processor of a general-purpose computer, special-purpose computer, or other programmable data processing device to produce a machine. The instructions, when executed through the computer processor or other programmable data processing device, allow the implementation of the functions / acts specified in the flowchart and / or block or blocks of the block diagram.Such processors may be, without limitation, general-purpose processors, special-purpose processors, application-specific processors, or programmable gate arrangements. Petition 870260008329, dated 01 / 28 / 2026, pp. 62 / 78 46 / 46 field.

[00130] The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of computer program systems, methods, and products according to various embodiments of this disclosure. In this respect, each block in the flowchart or block diagram may represent a module, segment, or portion of code, comprising one or more executable instructions to implement the specified logical function(s). It should also be noted that, in some alternative implementations, the functions observed in the block may occur out of the order shown in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in reverse order, depending on the functionality involved.It will also be noted that each block of the block diagrams and / or flowchart illustration, and combinations of blocks in the block diagrams and / or flowchart illustration, can be implemented by systems based on special-purpose hardware that perform the specified functions or actions, or combinations of special-purpose hardware and computer instructions.

[00131] Although the foregoing is directed to embodiments of the present disclosure, other and additional embodiments of the disclosure may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow. Petition 870260008329, dated 01 / 28 / 2026, pp. 63 / 78

Claims

1 / 4 CLAIMS 1. Method implemented by processor (102), CHARACTERIZED by comprising: receiving (910) an indication of a visual resource for modification and a natural language input prompt specifying a design objective to be applied to the indicated visual resource; identifying (920) visual components in the indicated visual resource and properties of the identified visual components to be modified based on the specified design objective; populating (930) one or more control panels with one or more controls to modify the identified visual components to achieve the specified design objective; receiving (940) an input in at least one of the one or more control panels; and modifying (950) the indicated visual resource based on the received input.

2. Method, according to claim 1, CHARACTERIZED in that the input comprises a slider control input associated with a design property of the visual feature.

3. Method, according to claim 2, CHARACTERIZED in that an adjustment to the slider input applies a global modification to the identified visual components.

4. A method according to claim 2, characterized in that the slider input is displayed in a first control panel, and in that an adjustment to the slider input modifies one or more of the properties of the identified visual components displayed in a second control panel.

5. Method, according to claim 2, CHARACTERIZED in that a range associated with the slider input comprises a minimum value associated with a first set of values ​​for the identified properties of the identified visual components and a maximum value associated with a second set of values ​​for the identified properties of the identified visual components, wherein the first set of values ​​and the second set of values ​​are identified based on the specified design objective.

6. Method, according to claim 1, CHARACTERIZED in that the input comprises an adjustment to a property of the identified properties associated with a visual component of the identified visual components.

7. Method, according to claim 1, CHARACTERIZED in that a range of valid values ​​for the identified properties of the identified visual components is associated with a defined amount of modification to be applied to the visual feature.

8. Method, according to claim 1, CHARACTERIZED in that a range of valid values ​​for the identified properties of the identified visual components is defined in relation to another visual component in a user interface.

9. Processing system, CHARACTERIZED by comprising: at least one memory (116) having executable instructions stored therein; and one or more processors (102) configured to execute the executable instructions so that the processing system: receives (910) an indication of a visual resource for modification and a natural language input prompt specifying a design objective to be applied to the indicated visual resource; identifies (920) visual components in the indicated visual resource and properties of the identified visual components to be modified based on the design objective Petition 870260008329, dated 01 / 28 / 2026, p.65 / 78 3 / 4 specified; fill (930) one or more control panels with one or more controls to modify the identified visual components to achieve the specified design objective; receive (940) an input in at least one of the one or more control panels; and modify (950) the indicated visual feature based on the received input.

10. Processing system according to claim 9, CHARACTERIZED in that the input comprises a slider control input associated with a design property of the visual feature.

11. Processing system according to claim 10, CHARACTERIZED in that an adjustment to the slider control input applies a global modification to the identified visual components.

12. Processing system, according to claim 10, CHARACTERIZED in that the slider control input is displayed on a first control panel, and in that an adjustment to the slider control input modifies one or more of the properties of the identified visual components displayed on a second control panel.

13. Processing system, according to claim 10, CHARACTERIZED in that a range associated with the slider control input comprises a minimum value associated with a first set of values ​​for the identified properties of the identified visual components and a maximum value associated with a second set of values ​​for the identified properties of the identified visual components, and wherein the first set of values ​​and the second set of values ​​are identified based on the specified design objective.

14. Processing system, according to Petition 870260008329, dated 01 / 28 / 2026, page 66 / 78 4 / 4 claim 9, CHARACTERIZED by the fact that the input comprises an adjustment to a property of the identified properties associated with a visual component of the identified visual components.

15. Processing system, according to claim 9, CHARACTERIZED in that a range of valid values ​​for the identified properties of the identified visual components is associated with a defined amount of modification to be applied to the visual feature.

16. Processing system, according to claim 9, CHARACTERIZED in that a range of valid values ​​for the identified properties of the identified visual components is defined in relation to another visual component in a user interface.

17. A computer-readable medium having executable instructions stored thereon which, when executed by one or more processors (102), performs an operation, CHARACTERIZED in that the operation comprises: receiving (910) an indication of a visual feature for modification and a natural language input prompt specifying a design objective to be applied to the indicated visual feature; identifying (920) visual components in the indicated visual feature and properties of the identified visual components to be modified based on the specified design objective; populating (930) one or more control panels with one or more controls to modify the identified visual components to achieve the specified design objective; receiving (940) an input in at least one of the one or more control panels; and modifying (950) the indicated visual feature based on the received input. Petition 870260008329, dated 01 / 28 / 2026, pp. 67 / 78