Systems, apparatus, and methods for material subgraph models

Abstract

Systems, apparatuses, and methods for material subgraph models are disclosed. The method may utilize an artificial intelligence (AI)-based material subgraph model including decomposition of a molecular structure of the material to generate subgraph graphics and implement subgraph modeling to extend the AI-based model for large and / or complex molecules and enhance material property prediction. Aspects may relate to utilizing enhanced material property prediction to improve overall performance of display-related products. For example, by using material property prediction with higher accuracy of sub-graph modeling, a display device may be manufactured using materials that are considered most appropriate and / or most effective. The method may include generating a subgraph comprising molecular substructures of the material, and applying an AI-based model to the subgraph to generate a material property prediction based on the molecular substructures of the material.

Description

[0001] Cross-references to related applications

[0002] This application claims the priority of U.S. Provisional Application No. 63 / 719,527, filed November 12, 2024, and U.S. Patent Application No. 19 / 182,427, filed April 17, 2025, the disclosures of which are incorporated herein by reference in their entirety as if fully set forth herein. Technical Field

[0003] Various aspects of some embodiments of this disclosure generally relate to machine learning and / or artificial intelligence. More specifically, the subject matter disclosed herein relates to determining material properties for displaying relevant products based on artificial intelligence. Background Technology

[0004] Material property prediction may involve assessing the physical, chemical, mechanical, and / or optical properties of materials that can be used to display related products. Therefore, it may be desirable to utilize computational models (e.g., AI-based models) to generate material property predictions, which can then be used to display various aspects of the design and / or manufacturing of related products, including material selection, performance optimization, and innovations in display technology.

[0005] The information disclosed in this background section is intended to enhance the understanding of the background of this disclosure, and therefore may contain information that does not constitute prior art. Summary of the Invention

[0006] This disclosure generally relates to electronic devices. More specifically, the subject matter disclosed herein relates to determining material properties for display-related electronic devices based on artificial intelligence.

[0007] Various aspects of some embodiments of this disclosure relate to material property prediction and / or analysis. For example, various aspects of some embodiments of this disclosure relate to improving the accuracy and efficiency of material property prediction by utilizing artificial intelligence and / or material subgraph models.

[0008] Material property prediction may involve assessing the physical, chemical, mechanical, and / or optical properties of materials that can be used to display related products. Therefore, it may be desirable to utilize computational models (e.g., AI-based models) to generate material property predictions, which can then be used to display various aspects of the design and / or manufacturing of related products, including material selection, performance optimization, and innovations in display technology. However, there may be issues associated with applying artificial intelligence (AI) techniques to material property prediction, such as maintaining the robustness and / or efficiency of AI and scaling to the relatively large size and complexity of the molecular structures of the materials that can be used to display related products.

[0009] Various aspects of some embodiments of this disclosure relate to systems and methods for AI-based material subgraph models, including: decomposition of the molecular structure of materials, generation of subgraph graphs, and implementation of subgraph modeling in a manner that allows the AI-based model to be extended to large and / or complex molecules and enhances its expressive power. Therefore, the disclosed embodiments can improve the overall performance of display-related products by utilizing materials deemed most suitable and / or most effective.

[0010] In some embodiments, a method for a material subgraph model includes: generating a subgraph by a processor, each subgraph in the subgraph including a molecular substructure of a material; applying an AI-based model to the subgraph by the processor to generate material property predictions based on the molecular substructure of the material; determining, by the processor, material-related functions for the production of a target device based on the material property predictions; and sending a signal by the processor to a component to control the component to perform the material-related functions for the production of the target device.

[0011] In some embodiments, the method may further include decomposing the molecular structure of the material into sub-molecular structures.

[0012] In some embodiments, decomposition may include breaking chemical bonds (BRIC) that have retrosynthetic significance.

[0013] In some embodiments, the method may further include generating a subgraph embedding based on the subgraph.

[0014] In some embodiments, generating the embedding of a subgraph may include processing the subgraph via a graph neural network.

[0015] In some embodiments, the method may further include subgraph-based embedding to generate subgraph graphs.

[0016] In some embodiments, the subgraph may include nodes and edges.

[0017] In some embodiments, each node in a node corresponds to a subgraph in a subgraph, and the value for each node in a node corresponds to an embedding in the embedding of the subgraph.

[0018] In some embodiments, each edge in the edge represents a relationship between subgraphs.

[0019] In some embodiments, the method may further include generating updated node embeddings based on the subgraph graph.

[0020] In some embodiments, generating updated node embeddings may include processing the subgraph graph via a graph neural network.

[0021] In some embodiments, AI-based models analyze subgraphs and model the relationships between subgraphs.

[0022] In some embodiments, the target device includes an organic light-emitting diode (OLED) display device.

[0023] In some embodiments, material property prediction is based on one or more of the following: physical properties of the material, chemical properties of the material, mechanical properties of the material, and optical properties of the material.

[0024] In some embodiments, an apparatus for a material subgraph model includes: one or more processors configured to perform: generating a subgraph, each subgraph including a molecular substructure of a material; applying an AI-based model to the subgraph to generate material property predictions based on the molecular substructure of the material; determining, based on the material property predictions, material-related functions for the production of a target apparatus; and sending signals to components to control the components to perform the material-related functions for the production of the target apparatus.

[0025] In some embodiments, one or more processors are further configured to perform: decomposing the molecular structure of a material into molecular substructures.

[0026] In some embodiments, one or more processors are further configured to perform: subgraph-based subgraph embedding.

[0027] In some embodiments, one or more processors are further configured to perform: subgraph-based embedding to generate subgraph graphs.

[0028] In some embodiments, one or more processors are further configured to perform: analyzing subgraph graphs using AI-based models and modeling relationships between subgraphs.

[0029] In some embodiments, a system for a material subgraph model includes: processing circuitry; and a memory storing instructions that, when executed by the processing circuitry, cause the processing circuitry to: generate a subgraph, each subgraph in which a material includes a molecular substructure; apply an AI-based model to the subgraph to generate material property predictions based on the molecular substructure of the material; determine, based on the material property predictions, material-related functions for the production of a target device; and send signals to components to control the components to perform the material-related functions for the production of the target device. Attached Figure Description

[0030] In the following sections, aspects of the subject matter disclosed herein will be described with reference to exemplary embodiments illustrated in the figures.

[0031] Figure 1This is a block diagram depicting a system (e.g., a factory) for producing products (e.g., electronic devices such as organic light-emitting diode (OLED) display devices) using an AI-based material subgraph model, according to some embodiments of the present disclosure.

[0032] Figure 2 This is a block diagram depicting an example of a computer device for predicting material properties, including a material subgraph modeling circuit, according to some embodiments of the present disclosure.

[0033] Figure 3 This is a block diagram depicting another example of a computer device for predicting material properties, including a material subgraph modeling circuit, according to some embodiments of the present disclosure.

[0034] Figure 4 This describes some embodiments according to the present disclosure. Figure 2 A diagram illustrating an example process for decomposing the molecular structure of materials, implemented using a material subgraph modeling circuit.

[0035] Figure 5 This describes some embodiments according to the present disclosure. Figure 2 A diagram illustrating an example process for generating embedded subgraphs, implemented using a material subgraph modeling circuit.

[0036] Figure 6 This describes some embodiments according to the present disclosure. Figure 2 A diagram illustrating an example process for generating subgraph graphics, implemented using a material subgraph modeling circuit.

[0037] Figure 7A This describes some embodiments according to the present disclosure. Figure 3 A diagram illustrating an example process for generating updated subgraphs using an embedded material subgraph modeling circuit.

[0038] Figure 7B It is an embedding of the generated updated subgraph according to some embodiments of the present disclosure and is described by Figure 3 A diagram showing an example configuration of a converter implemented using a material subgraph modeling circuit.

[0039] Figure 8 This is a block diagram of an electronic device in a network environment according to some embodiments of the present disclosure.

[0040] Figure 9 This is a flowchart depicting example operations of a method for predicting material properties using an AI-based material subgraph model, according to some embodiments of the present disclosure. Detailed Implementation

[0041] In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of this disclosure. However, those skilled in the art will understand that the disclosed aspects can be practiced without these specific details. In other instances, well-known methods, procedures, components, and circuits have not been described in detail so as not to obscure the subject matter disclosed herein.

[0042] Throughout this specification, references to "an embodiment" or "an embodiment" mean that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment disclosed herein. Therefore, the phrases "in one embodiment," "in an embodiment," or "according to an embodiment" (or other phrases with similar meanings) appearing throughout this specification may not necessarily refer to the same embodiment. Furthermore, in some embodiments (e.g., in one or more embodiments), particular features, structures, or characteristics may be combined in any suitable manner. In this regard, as used herein, the word "exemplary" means "serving as an example, instance, or illustration." Any embodiment described herein as "exemplary" should not be construed as necessarily superior to or advantageous to other embodiments. Additionally, particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. Furthermore, depending on the context discussed herein, singular terms may include corresponding plural forms, and plural terms may include corresponding singular forms. Similarly, hyphenated terms (e.g., "two-dimensional", "pre-determined", "pixel-specific", etc.) are occasionally used interchangeably with their non-hyphenated versions (e.g., "two-dimensional", "pre-determined", "pixel specific", etc.), and uppercase entries (e.g., "counter clock", "row select", "pixout", etc.) are interchangeable with their non-uppercase versions (e.g., "counter clock", "row select", "pixout", etc.). This occasional interchangeability should not be considered inconsistent with each other.

[0043] Additionally, depending on the context discussed herein, singular terms may include their corresponding plural forms, and plural terms may include their corresponding singular forms. Further note that the various figures (including component figures) shown and discussed herein are for illustrative purposes only and are not drawn to scale. For example, for clarity, some dimensions of the components may be enlarged relative to other components. Furthermore, reference numerals have been repeated in the figures where appropriate to indicate corresponding and / or similar components.

[0044] The terminology used herein is for the purpose of describing some exemplary embodiments only and is not intended to limit the claimed subject matter. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that, when used in this specification, the terms “comprising” and / or “including” specify the presence of the stated features, integers, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0045] It will be understood that when an element or layer is referred to as "on" another element or layer, "connected to," or "attached to" another element or layer, it may be directly on, directly connected to, or directly attached to the other element or layer, or an intermediary element or layer may be present. In contrast, when an element is referred to as "directly on" another element or layer, "directly connected to," or "directly attached to" another element or layer, no intermediary element or layer is present. Throughout the text, the same reference numerals refer to the same elements. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0046] As used herein, the terms “first,” “second,” etc., serve as labels for the preceding nouns and do not imply any type of ordering (e.g., spatial, temporal, logical, etc.) unless explicitly stated otherwise. Furthermore, the same reference numerals may be used in two or more figures to refer to parts, components, blocks, circuits, units, or modules having the same or similar functions. However, this usage is solely for the purpose of brevity and ease of discussion; it does not imply that the construction or architectural details of such parts or units are identical in all embodiments, or that these commonly referenced parts / modules are the only way to implement some of the exemplary embodiments disclosed herein.

[0047] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the subject matter pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having a meaning consistent with their meaning in the context of the relevant field, and shall not be interpreted in an idealized or overly formal sense unless expressly stated otherwise herein.

[0048] As used herein, the term "module" means any combination of software, firmware, and / or hardware configured to provide the functionality described herein in conjunction with the module description. For example, software may be implemented as a software package, code, and / or instruction set or instructions, and as used in any implementation described herein, the term "hardware" may include, for example, individual or combined components, hardware circuitry, programmable circuitry, state machine circuitry, and / or firmware storing instructions executed by programmable circuitry. Modules may be implemented collectively or individually as circuitry forming part of a larger system, such as, but not limited to, integrated circuits (ICs), system-on-a-chip (SoCs), and components.

[0049] In recent years, the display industry has focused on developing cutting-edge, next-generation display materials for products. The goal may be to achieve new, highly efficient, and low-cost display-related products. A key process for achieving improved products may be the efficient and precise acquisition of material properties. Effective determination of the properties and / or characteristics of potential materials used in product manufacturing can accelerate the development and optimization of new materials.

[0050] Density functional theory (DFT) simulations can be used to obtain material properties based on physical constraints prior to any experimental verification. However, the accuracy and / or reliability of predictions of material properties provided by DFT simulations (and other traditional mechanisms) can be limited. For example, DFT methods may experience a delay (a relatively long time to reach a steady-state solution). The low computational efficiency of some current material property analysis mechanisms, such as DFT, may limit their application in the rapid discovery of new materials.

[0051] Additionally, there may be some AI-based methods currently used to provide automated material property analysis and / or prediction tools. However, these AI-based models may fail to effectively capture the diverse relationships that can be associated with the complex molecular structure of materials, leading to the loss of crucial relational information.

[0052] These AI-based models may be ill-suited for capturing long-range interactions, which could limit their expressive power and lead to performance degradation when analyzing macromolecules. Furthermore, the sampling methods used for these AI-based models may rely on conventional modeling features (e.g., node and / or edge removal), which may not be well-suited for utilizing chemical decomposition techniques and thus may reduce the model's ability to effectively handle macromolecules (e.g., molecules with hundreds of atoms), where many materials used to visualize related devices may comprise macromolecules. In some cases, limitations in material property analysis and / or prediction may stem from reliance on publicly available datasets that may contain relatively small molecules (e.g., composed of only tens of atoms).

[0053] To improve material property analysis and / or prediction, and consequently enhance display-related products, embodiments implement functionalities related to the generation, training, and utilization of AI-based subgraph models. These AI-based subgraph models can be extended to large and complex molecules and have enhanced expressive power to improve the accuracy and / or performance of material property prediction. As mentioned above, materials used to produce display-related devices can often involve macromolecules, and therefore, AI-based models capable of handling such complexity may be required. As disclosed herein, AI-based subgraph models achieve extension to macromolecules by utilizing subgraphs. For example, AI-based subgraph models can utilize subgraph decomposition of the molecular structure of materials (e.g., chemistry) instead of relying on more conventional AI modeling techniques (e.g., edge removal, node removal, etc.).

[0054] As disclosed herein, AI-based subgraph models can also achieve enhanced expressive power. By enhancing the expressive power of AI-based subgraph models (e.g., effectively modeling long-range interactions by modeling relationships between subgraphs), these models can achieve better performance in predicting material properties.

[0055] Various aspects of some embodiments of this disclosure provide AI-based subgraph models to alleviate (e.g., overcome) the aforementioned problems by implementing the following functionalities: subgraph selection, which may involve identifying and / or selecting meaningful substructures within a molecular graph based on specific chemical principles (e.g., breaking chemical bonds with retrosynthetic significance (BRIC), hierarchical decomposition strategies, etc.) using a domain-specific subgraph sampler; subgraph representation, which may include representing each subgraph as a corresponding node within a constructed graph (referred to herein as a subgraph graph), wherein edges between nodes may represent interactions or relationships between corresponding subgraphs; interaction modeling, which may involve capturing and / or modeling interactions (and dependencies) between subgraphs within a subgraph graph to preserve relational information and achieve representation of complex structural patterns; and material property prediction, which may involve applying neural networks to the subgraph graph to predict material properties based on aggregated and interacting subgraph representations.

[0056] Figure 1 This is a block diagram depicting a system 100 (e.g., a factory) for producing display-related products (e.g., electronic devices such as organic light-emitting diode (OLED) display devices) according to some embodiments of the present disclosure.

[0057] Manufacturing products (for example, in a factory or production line) may include various processes to ensure that specific quality standards are met. In some embodiments, Figure 1 System 100 (e.g., a factory) can produce products that may be referred to as target devices in this disclosure (e.g., electronic devices such as display devices and / or integrated circuits). Figure 1 As seen, system 100 may include production line 108. Production line 108 may include machines, machinery and / or apparatus that take raw materials 106 and / or components as input and assemble, build and / or produce one or more products such as display devices (e.g., OLED display devices).

[0058] When manufacturing target devices such as display-related products (e.g., OLED display devices), material property prediction can be used to identify and select the most suitable organic materials for each layer of the target device. Production design system 102 may include material property prediction system 103, which may be implemented as a computer device capable of generating, training, and / or utilizing material subgraph model 120. Material subgraph model 120 may be an artificial intelligence (AI) and / or machine learning (ML) based model (also referred to herein as a “neural network”), which can be generated, trained, and / or utilized according to the subgraph modeling functions disclosed herein, and can achieve improved efficiency and / or accuracy in predicting material properties. Production design system 102 can then utilize the enhanced accuracy and / or predictive performance of material property prediction system 103 for its functions related to the design, testing, and / or production of display-related products.

[0059] Material subgraph model 120 can be used to implement computational modeling techniques that can generate material property predictions based on the physical, chemical, electrical, mechanical, and / or optical properties of materials (e.g., the molecular structure of potential materials). Subsequently, based on the enhanced material property predictions implemented by material property prediction system 103 and / or material subgraph model 120, production design system 102 can be able to perform multiple functions involving one or more of raw materials 106 (e.g., selection, design, manufacturing, and / or verification, etc.), one or more of raw materials 106 can be considered appropriate and / or optimal for designing (e.g., computer-aided design) products, which can ultimately be used for the actual manufacturing of products in production line 108.

[0060] The production design system 102 may have the capability to perform aspects related to the automated simulation, design, manufacturing, and / or verification of raw material 106, such as one or more different organic materials for each layer of an OLED display device. As an operational example, the production design system 102 may utilize the material subgraph model 120 in computational predictions to identify and / or analyze key properties for raw material 106, such as luminous efficiency, stability, and charge transport properties prior to the subsequent synthesis, testing, and / or utilization (e.g., for manufacturing a product) of raw material 106. For example, the production design system 102 may utilize (e.g., generated by the material property prediction system 103 and / or the material subgraph model 120) enhanced material property predictions to select (e.g., molecules that can meet the defined criteria for raw material 106) and / or filter (e.g., molecules that may not meet the defined criteria for raw material 106) raw material 106 and / or the molecular structure of raw material 106 for subsequent design, manufacturing, verification, and / or utilization of raw material 106.

[0061] In another example, the production design system 102 can control and / or execute functions related to the synthesis of feedstock 106. For example, the production design system 102 can control one or more automated functions for a chemical reactor to synthesize feedstock 106 using molecules selected based on enhanced material property predictions (e.g., generated by the material property prediction system 103 and / or the material subgraph model 120). The production design system 102 can control and / or execute multiple functions related to the material fabrication and / or synthesis of feedstock 106. Thus, enhanced material property predictions can be utilized in the fabrication and / or synthesis of feedstock 106 to improve the speed, efficiency, and performance of the process (e.g., to improve the selection of candidate molecules and / or materials, etc.), which can reduce overall time consumption (e.g., the delay between molecular design and validation of feedstock 106) and can optimize the design and / or performance of the synthesized feedstock 106.

[0062] In another operational example, the production design system 102 can control and / or perform automated testing and / or verification functions involving raw material 106 based on enhanced material property predictions (e.g., generated by the material property prediction system 103 and / or the material subgraph model 120). For example, the production design system 102 can control automated testing of the synthesized raw material 106 (e.g., OLED test material), which may involve analyzing light emitted from a product using raw material 106 to measure operational properties (e.g., electroluminescence), thereby verifying and / or confirming (e.g., comparing the measured properties of the tested material with theoretical and / or predetermined properties) the performance of raw material 106. Thus, enhanced material property predictions can be utilized in the verification and / or testing of raw material 106 to ensure that the manufactured and / or synthesized raw material 106 (and the products manufactured using raw material 106) meets desired quality and / or standards (e.g., color and / or brightness, etc.) before being utilized downstream in system 100 (e.g., in production line 108 for manufacturing display-related products) (e.g., the synthesized material is processed into a film for layers).

[0063] In another operational example, production design system 102 can control and / or execute functions related to the manufacture of display-related devices downstream of system 100 (e.g., in production line 108) after testing and / or validating raw material 106. For example, production design system 102 can transmit (e.g., send) commands (e.g., control signals) to systems in production line 108 (such as manufacturing systems and / or processing machines) to control the automated production of OLED display devices using materials selected based on enhanced material property predictions (e.g., materials selected from raw material 106) generated by (e.g., by material property prediction system 103 and / or material subgraph model 120), thereby optimizing overall display performance and minimizing development time and cost.

[0064] Figure 2 This is a block diagram depicting a computer device 200 for predicting material properties, including a material subgraph modeling circuit 250 for implementing subgraph modeling, according to some embodiments of the present disclosure.

[0065] like Figure 2 As illustrated herein, computer device 200 (e.g., one or more computers and / or one or more computer systems) may include memory 211 (e.g., memory and / or storage devices), processor 212, and material subgraph modeling circuitry 250 configured for implementing subgraph modeling. Memory 211 may be connected to… Figure 8 The memory 830 corresponds to this. The processor 212 can be connected to... Figure 8 Corresponding to processor 820. As a general description, as disclosed herein, material subgraph modeling circuit 250 can perform one or more functions related to subgraph modeling, which may involve performing material (e.g., chemical) subgraph sampling, obtaining subgraph embeddings, generating subgraph graphs, modeling interactions between subgraphs, and generating predictions of material properties based on subgraphs.

[0066] According to some embodiments, the computer device 200 can utilize an AI-based model (e.g., material subgraph model 120) generated, trained, and / or utilized by the material subgraph modeling circuit 250 and related functions to achieve improved accuracy, efficiency, and / or performance. For example, as disclosed herein, the material subgraph modeling circuit 250 can be configured to perform one or more of the related functions during the training and / or inference phases of the AI-based model (e.g., material subgraph model 120).

[0067] Computer device 200 may include a computer system capable of implementing AI-related functions, including model training, computation, inference, and various AI-based applications. For example, computer device 200 may be implemented as, but is not limited to, a desktop PC, laptop computer, smartphone, tablet PC, and / or server. Computer device 200 may include a system in which a cloud computing environment is established. However, some embodiments are not limited thereto. As described herein, computer device 200 may be implemented as any system, apparatus, or device capable of implementing AI-based applications and functions (e.g., material property prediction, subgraph modeling, etc.), such as (e.g., Figure 1 (As shown in the diagram) Material property prediction system 103. In some embodiments, as disclosed herein, computer device 200 can perform various functions related to the manufacture and / or inspection of OLED display devices (e.g., material selection, product design, etc.), such as (e.g., Figure 1(As shown in the diagram) Production design system 102. Computer device 200 may include one or more processors for performing one or more processes of this disclosure.

[0068] In some embodiments, as disclosed herein, memory 211 may store data and / or AI models, including material subgraph model 120, associated with AI-based applications such as material property prediction, subgraph modeling, etc. In some embodiments, memory 211 may store models generated, trained, and / or utilized by material subgraph modeling circuitry 250. In some embodiments, as disclosed herein, memory 211 may store material property predictions generated by material subgraph modeling circuitry 250 and utilize subgraph modeling capabilities.

[0069] In some embodiments, as disclosed herein, processor 212 may include various processing circuitry and may control the overall operation of computer device 200, including AI-based applications supported by AI models (e.g., material subgraph model 120) generated, trained, and utilized by material subgraph modeling circuitry 250. In some embodiments, as disclosed herein, processor 212 may include various processing circuitry (e.g., one or more processing circuits) and may control the overall operation of computer device 200 (e.g., computer system), including AI-based applications supported by material property predictions 236 generated by material subgraph modeling circuitry 250. In some embodiments, processor 212 may be implemented as, for example, a digital signal processor (DSP), a microprocessor, or a time controller (TCON), but is not limited thereto. Processor 212 may be, for example, and not limited to, one or more of a dedicated processor, central processing unit (CPU), microcontroller unit (MCU), microprocessor unit (MPU), controller, application processor (AP), communication processor (CP), and ARM processor, or may be defined as one of the foregoing terms. Alternatively, the processor 212 can be implemented as a system-on-a-chip (SoC) in which processing algorithms are provided, or it can be implemented in the form of a field-programmable gate array (FPGA), but is not limited thereto.

[0070] The material subgraph modeling circuit 250 can be implemented using any suitable hardware, firmware (e.g., application-specific integrated circuit), software, or a combination of software, firmware, and hardware. As described in more detail herein, the material subgraph modeling circuit 250 can be configured to perform material subgraph sampling. Material subgraph sampling can utilize chemical decomposition techniques such as BRIC and / or hierarchical decomposition strategies to extract material-related molecular substructures. Therefore, the material subgraph modeling circuit 250 can be configured to perform the extraction of functional groups and / or structural motifs that may be crucial for determining material properties.

[0071] As described in more detail herein, the material subgraph modeling circuit 250 can be configured to render subgraphs that provide modular (e.g., having multiple segments and / or parts) and graph-based representations of properties. Subgraphs can involve generating new graphs where selected subgraphs can be used as nodes and edges can represent their interactions. Therefore, the material subgraph modeling circuit 250 can model long-term dependencies between subgraphs, which can support accurate prediction of material properties.

[0072] As described in more detail herein, the material subgraph modeling circuit 250 can be configured to model the interactions between subgraphs. Therefore, the material subgraph modeling circuit 250 can capture dependencies and / or interactions between subgraphs by leveraging a graph-based representation of the subgraphs, thereby preserving relational information and achieving improved accuracy in material property prediction by maintaining relational interactions. By modeling the interactions between subgraphs, the material subgraph modeling circuit 250 can provide AI models with an enhanced level of expression (e.g., greater than the expression level of the 3-Weisfeiler-Lehman (3-WL) test), enabling AI models to distinguish complex and large-scale graph structures in a way that can improve upon the capabilities of simpler models.

[0073] Figure 2 The illustrated material subgraph modeling circuit 250 can be configured to receive input 201, which may include data representing the molecular structure of a material (e.g., a chemical) relevant to the production and / or manufacturing of a product. In some embodiments, input 201 may be an initial graphical representation of the molecular structure of the material, wherein the chemical composition of the material may include large and complex molecules (e.g., having hundreds of atoms). The material subgraph modeling circuit 250 can utilize input 201 to perform material subgraph sampling. For example, the material subgraph modeling circuit 250 can perform decomposition on the molecular structure represented in input 201 to extract one or more molecular substructures, wherein the molecular substructures may be further represented as subgraph 202. In some embodiments, the material subgraph modeling circuit 250 can be configured to implement BRIC for decomposing input 201. The material subgraph modeling circuit 250 can perform hierarchical decomposition according to BRIC to identify molecular substructures within the molecules of the material. The decomposition may be based on a determined number (e.g., minimum, maximum, etc.) of functional groups and / or structural motifs to be extracted.

[0074] In some embodiments, the material subgraph modeling circuit 250 may be configured to control and / or determine the number of subgraphs 202 generated from molecular substructures due to decomposition. For example, the material subgraph modeling circuit 250 may generate a determined number of subgraphs 202 such that the aggregate composition of the subgraphs 202 can overwrite the original graph represented by input 201. The number of subgraphs 202 generated from the decomposition may be determined based on application-related factors, including but not limited to: the computational resources and / or configuration of the material property prediction system 103; and / or efficiency and / or performance-related metrics (e.g., maintaining computational speed), etc. In some embodiments, the material subgraph modeling circuit 250 may be configured to prioritize one or more molecular substructures selected during decomposition for extraction. For example, a specific substructure of a molecule may be considered to have chemical significance to the molecular structure of the material, and therefore may be extracted by the material subgraph modeling circuit 250 prior to other molecular substructures during decomposition and represented in the subgraph 202. Embodiments are not limited thereto, and the material subgraph modeling circuit 250 may be configured to utilize other material and / or molecular decomposition mechanisms considered appropriate and / or suitable.

[0075] Figure 4 Examples of a decomposition process 400 of a subgraph 202 generating molecular substructures implemented by a material subgraph modeling circuit 250 according to some embodiments of the present disclosure are described.

[0076] In some embodiments, the material subgraph modeling circuit 250 may receive input 201, which includes data representing the structure of molecules 405 related to the material used in the production of the product. For example... Figure 4 The molecule 405 illustrated in the figure may have a structure that forms a compound comprising a group of atoms bonded together (e.g., NH2, NO). The decomposition process 400 for molecule 405 may involve systematically fragmenting molecule 405 by identifying portions and / or regions (or subregions) and / or breaking bonds considered appropriate (e.g., breaking bonds with retrosynthetic significance, etc.). The decomposition process 400 illustrated in the figure may involve decomposing molecule 405 based on the identification of ring 410, acyclic moiety 411, and carbon-carbon single bond 412 comprising molecule 405. Ring 410, acyclic moiety 411, and carbon-carbon single bond 412 can be extracted from the decomposition of molecule 405 such that molecular substructures 421 to 426 of molecule 405 can be formed as fragments (e.g., constructed by arranging one or more of the extracted portions of molecule 405), which may be functional groups and / or structural motifs relevant to determining the property chemistry of the material. The molecular substructures 421 to 426 formed from the complex structure of decomposed molecule 405 can be represented as subgraph 202. In Figure 4In the example, each of the molecular substructures 421 to 426 can be represented as a different corresponding subgraph 202.

[0077] Refer again Figure 2 The material subgraph modeling circuit 250 can be configured to obtain subgraph embeddings (also referred to as subgraph embeddings) 210 based on the subgraph 202. For example, the material subgraph modeling circuit 250 can input the data subgraph 202 (representing molecular substructures) into a graph neural network (GNN) 215 to generate subgraph embeddings 210, wherein each of the subgraph embeddings 210 can be a vector representation of a portion of the properties of the molecule (input 201) from the corresponding subgraph 202.

[0078] GNN 215 can process AI-based graph-structured data, such as subgraphs 202 including data representing molecular substructures. As an example, each subgraph 202 can be structured as a set of nodes and edges, and GNN 215 can analyze the subgraph 202 to capture features, properties, and / or encodings of data (e.g., nodes) and relationships (e.g., edges) involved in the learned subgraph embeddings 210. The subgraph embeddings 210 can be obtained in a format that the machine learning model can understand and use for downstream tasks. The functions performed by GNN 215 to obtain the subgraph embeddings 210 can involve tokenization, embedding, and / or positional encoding. In some embodiments, GNN 215 can receive the subgraph 202 and iteratively obtain the subgraph embeddings 210 as a function of the training process and / or during inference for the machine learning model performed by processor 212.

[0079] In some embodiments, the material subgraph modeling circuit 250 can be configured to generate a subgraph graph 220 as another representation of the subgraph 202. Figure 2 The illustration shows that subgraph 220 can be a graph structure that represents data as a set of nodes 221 and edges 222, wherein each of the nodes 221 within the constructed subgraph 220 can represent a corresponding subgraph 202, and the edges 222 (between nodes 221) can represent the interactions and / or relationships between the corresponding subgraphs 202. Each node 221 can have associated features and / or properties describing its attributes, and the edges 222 connecting the nodes 221 can represent the relationships between them. Depending on the data, the edges 222 can be directed (e.g., information flows in one direction) and / or undirected (e.g., information flows in both directions).

[0080] Figure 5 Examples of processes for generating subgraph graphics 220, implemented by a material subgraph modeling circuit 250 according to some embodiments of the present disclosure, are described.

[0081] Figure 5The process may involve constructing the subgraph graph 220 into another graph-based representation of subgraphs 202a to 202f (including molecular substructures). As previously described, each node 221a to 221f of the subgraph graph 220 may represent a corresponding one of subgraphs 202a to 202f, edges 222a to 222g may indicate interactions and / or relationships between subgraphs 202a to 202f, and the values ​​for each of nodes 221a to 221f may correspond to subgraph embeddings 210a to 210f (subgraphs for node representation). In the example of subgraph graph 220, node 221a can represent subgraph 202a, and the value of node 221a can correspond to subgraph embedding 210a; node 221b can represent subgraph 202b, and the value of node 221b can correspond to subgraph embedding 210b; node 221c can represent subgraph 202c, and the value of node 221c can correspond to subgraph embedding 210c, and so on. The structure of subgraph graph 220 can be represented as adjacency matrix 510, where each cell of adjacency matrix 510 can indicate the structural properties of subgraph graph 220 (e.g., whether there is an edge between two nodes, etc.).

[0082] In some embodiments, the process may involve calculating the relationships and / or interactions between subgraphs 202a to 202f in a manner that models the relationships through hierarchical relationships. Distance methods such as Jaccard distance and / or fingerprints may be used to calculate the relationships between subgraphs 202a to 202f, and the relationship dependencies may be graphically represented in the structure of the subgraph graph 220 (e.g., the number of connections and / or edges, distances between nodes, etc.). For example, edge 222a may represent a relationship dependency between node 221a (representing subgraph 202a) and node 221c (representing subgraph 202c), and edge 222e may represent a relationship dependency between node 221c (representing subgraph 202c) and node 221e (representing subgraph 202e). Accordingly, the process may construct the subgraph graph 220 such that it provides a graph-based representation that models the relationships between molecular substructures (included in subgraphs 202a to 202f). In some embodiments, when three-dimensional (3D) coordinates are provided as input, interactions between molecular structures can also be modeled using coordinate distances in subgraph graph 220. Subgraph graph 220 can be used to capture and model interactions and / or dependencies between subgraphs 202 in a manner that preserves relational information and enables representations of complex structural patterns, such as the complex molecular structures of materials used in products.

[0083] In some embodiments, the material subgraph modeling circuit 250 may be configured to implement another graph neural network 230 in order to process data in the subgraph graph 220. Figure 6 Examples are depicted of a process for updating a subgraph graph 220, implemented by a material subgraph modeling circuit 250 according to some embodiments of the present disclosure. Accordingly, the material subgraph modeling circuit 250 can update the subgraph embedding (represented in the subgraph graph 220) by utilizing a graph neural network 230.

[0084] Figure 6 The diagram illustrates that a graph neural network 230 can iteratively update the graph embeddings of nodes 221 by aggregating information from neighboring nodes 221 within a subgraph graph 220. Therefore, a material subgraph modeling circuit 250 performs the transfer of information from a subgraph (e.g., node 221) to neighboring subgraphs (e.g., neighboring nodes 221). After determining the number of iterations (layers), the graph neural network 230 can generate a subgraph graph 220 with updated node embeddings 605. The updated node embeddings 605 can be aggregated by the graph neural network 230, which can generate a single vector 231 representing the aggregation of the updated node embeddings 605. In some embodiments, the embeddings of nodes 221a to 221f in the initially constructed subgraph graph 220 can be utilized (e.g., before the update). In some embodiments, the final node and / or graph representations obtained by the graph neural network 230 from analyzing the subgraph graph 220 can be used to generate a material property prediction 236 as output. Figure 6 In the example process, vector 231 can be passed to a multilayer perceptron (MLP) function 235. MLP function 235 can be a transformation function that processes (e.g., with updated node embeddings 605) the final node and / or graph representation of subgraph graph 220 for further processing tasks (such as classification and / or regression), and outputs a material property prediction 236. As previously described, the material subgraph modeling circuit 250 can model long-range interactions in a manner that increases expressiveness (e.g., greater than 3-WL) by modeling relationships between subgraphs, and thus can increase the accuracy and / or performance of the material property prediction 236.

[0085] Figure 3 This is a block diagram depicting another configuration of a computer device 300 for predicting material properties, including a material subgraph modeling circuit 310 implementing subgraph modeling, according to some embodiments of the present disclosure.

[0086] The computer device 300 and the material subgraph modeling circuit 310 can be compared with... Figure 2 The described computer device 200 and material subgraph modeling circuit 250 have substantially similar configurations and / or functions. For the sake of brevity, see references to... Figure 3The differentiated components and / or functions of the material subgraph modeling circuit 310 according to some embodiments are described in detail.

[0087] The material subgraph modeling circuit 310 can be configured to implement the transformer 305 to learn the relationships between subgraphs 202. For example, the material subgraph modeling circuit 310 can utilize the transformer 305 to process the subgraph embedding 210 and generate updated embeddings from the subgraph embedding 210 (e.g., without generating subgraph graph 220).

[0088] Figure 7A Examples of processes for generating updated embeddings 705, implemented by a material subgraph modeling circuit 310 according to some embodiments of the present disclosure, are described. Transformer 305 can be configured to model nonlinear interdependencies between markers, and therefore can be used to model interdependencies between subgraphs represented by subgraph embeddings 210. Transformer 305 can generate updated embeddings 705 by passing subgraph embeddings 210 through a series of layers, wherein each embedding can be refined based on its context within the sequence, generating context-aware embeddings that capture the relationships between subgraphs, thereby modeling the relationships between molecular substructures.

[0089] The updated embeddings 705 (for multiple subgraphs) can be aggregated, which can generate a single vector 231 (representing the aggregation of the updated embeddings 705). In some embodiments, the results for the updated embeddings 705 (e.g., calculations of the last layer of transformer 305) can be used to generate a material property prediction 236 as output. Vector 231 can be passed to an MLP function 235, which processes the results from the updated embeddings 705 for further processing tasks (such as classification and / or degradation), and outputs the material property prediction 236.

[0090] Figure 7BExamples of circuitry that can be used to implement the transformer functionality in transformer 305 are depicted. The circuitry of transformer 305 may include: a multi-head attention circuit 706 (denoted as multi-head attention), which can be configured to execute multiple attention mechanisms in parallel to process information from the input sequence and then cascade and / or linearly transform the dependent attention outputs into the desired dimension; addition and normalization circuits 707, 709 (denoted as addition and normalization), which can be configured to perform a combination operation, wherein the output of a layer is added to its original input (e.g., addition) and then a layer normalization step (“normalization”) is subsequently performed; and a feedforward circuit 708 (denoted as feedforward circuit), which can be configured to pass information in one direction. Therefore, the material subgraph modeling circuit 310 can leverage the relational modeling capabilities of transformer 305 in a way that improves accuracy in material property prediction 236 by preserving key interactions.

[0091] Figure 8 This is a block diagram of an electronic device in a network environment 800 according to some embodiments of the present disclosure.

[0092] refer to Figure 8 In network environment 800, electronic device 801 can communicate with external electronic device 802 via a first network 898 (e.g., a short-range wireless communication network), or with external electronic device 804 or server 808 via a second network 899 (e.g., a long-range wireless communication network). Electronic device 801 can communicate with external electronic device 804 via server 808. Electronic device 801 may include processor 820, memory 830, input device 850, sound output device 855, display device 860, audio module 870, sensor module 876, interface 877, haptic module 879, camera module 880, power management module 888, battery 889, communication module 890, subscriber identification module (SIM) 896, and / or antenna module 897. In one embodiment, at least one of the components (e.g., display device 860 or camera module 880) may be omitted from electronic device 801, or one or more other components may be added to electronic device 801. Some of the components may be implemented as a single integrated circuit (IC). For example, a sensor module 876 (e.g., a fingerprint sensor, an iris sensor, or an illuminance sensor) may be embedded in a display device 860.

[0093] The processor 820 can execute software (e.g., program 840) to control at least one other component (e.g., hardware or software component) of the electronic device 801 connected to the processor 820, and can perform various data processing or calculations.

[0094] As at least part of data processing or computation, processor 820 can load commands or data received from another component (e.g., sensor module 876 or communication module 890) into volatile memory 832, process the commands or data stored in volatile memory 832, and store the resulting data in non-volatile memory 834. Processor 820 may include a main processor 821 (e.g., a central processing unit or application processor (AP)) and an auxiliary processor 823 (e.g., a graphics processing unit (GPU), image signal processor (ISP), sensor hub processor, or communication processor (CP)) that may operate independently of or in conjunction with the main processor 821. Alternatively, auxiliary processor 823 may be adapted to consume less power than the main processor 821 or to perform specific functions. Auxiliary processor 823 may be implemented separately from or as part of the main processor 821.

[0095] The auxiliary processor 823 can control at least some of the functions or states associated with at least one component (e.g., display device 860, sensor module 876, or communication module 890), rather than the main processor 821 controlling it while the main processor 821 is in an inactive (e.g., sleep) state, or together with the main processor 821 controlling it while the main processor 821 is in an active state (e.g., executing an application). The auxiliary processor 823 (e.g., image signal processor or communication processor) can be implemented as part of another component (e.g., camera module 880 or communication module 890) functionally associated with the auxiliary processor 823.

[0096] Memory 830 may store various data used by at least one component of electronic device 801 (e.g., processor 820 or sensor module 876). The various data may include, for example, software (e.g., program 840) and input or output data for commands associated with it. Memory 830 may include volatile memory 832 or non-volatile memory 834. Non-volatile memory 834 may include internal memory 836 and / or external memory 838.

[0097] Program 840 can be stored as software in memory 830 and may include, for example, an operating system (OS) 842, middleware 844, or application program 846.

[0098] Input device 850 can receive commands or data used by another component of electronic device 801 (e.g., processor 820) from outside electronic device 801 (e.g., a user). Input device 850 may include, for example, a microphone, mouse, or keyboard.

[0099] The sound output device 855 can output sound signals to the outside of the electronic device 801. The sound output device 855 may include, for example, a speaker or a receiver. The speaker can be used for general purposes such as playing multimedia or recording, and the receiver can be used to receive incoming calls. The receiver can be implemented separately from the speaker or implemented as part of the speaker.

[0100] Display device 860 can visually provide information to the outside of electronic device 801 (e.g., to a user). Display device 860 may include, for example, a display, a holographic device, or a projector, and may include control circuitry to control a corresponding one of the display, holographic device, and projector. Display device 860 may include touch circuitry adapted to detect touch, or may include sensor circuitry (e.g., a pressure sensor) adapted to measure the intensity of the force caused by touch.

[0101] The audio module 870 can convert sound into electrical signals and vice versa. The audio module 870 can obtain sound via the input device 850, or can output sound via the sound output device 855 or via headphones of the external electronic device 802 directly (e.g., wired) or wirelessly connected to the electronic device 801.

[0102] Sensor module 876 can detect the operating state of electronic device 801 (e.g., power or temperature) or the environmental state outside electronic device 801 (e.g., the user's state). Sensor module 876 can then generate an electrical signal or data value corresponding to the detected state. Sensor module 876 may include, for example, a gesture sensor, gyroscope sensor, barometric pressure sensor, magnetic sensor, accelerometer, grip sensor, proximity sensor, color sensor, infrared (IR) sensor, biometric sensor, temperature sensor, humidity sensor, and / or illuminance sensor.

[0103] Interface 877 may support one or more specified protocols for direct (e.g., wired) or wireless connection of electronic device 801 to external electronic device 802. Interface 877 may include, for example, a High Definition Multimedia Interface (HDMI), a Universal Serial Bus (USB) interface, a Secure Digital Card (SD) interface, or an audio interface.

[0104] Connection terminal 878 may include a connector through which electronic device 801 can be physically connected to external electronic device 802. Connection terminal 878 may include, for example, an HDMI connector, a USB connector, an SD card connector, or an audio connector (e.g., a headphone connector).

[0105] The tactile module 879 can convert electrical signals into mechanical stimulation (e.g., vibration or movement) or electrical stimulation, which can be recognized by the user through touch or kinesthesia. The tactile module 879 may include, for example, a motor, a piezoelectric element, or an electrical stimulator.

[0106] Camera module 880 can capture still or moving images. Camera module 880 may include one or more lenses, an image sensor, an image signal processor, or a flash. Power management module 888 can manage the power supplied to electronic device 801. Power management module 888 may be implemented as at least a part of, for example, a power management integrated circuit (PMIC).

[0107] Battery 889 can supply power to at least one component of electronic device 801. Battery 889 may include, for example, a non-rechargeable primary battery, a rechargeable secondary battery, or a fuel cell.

[0108] Communication module 890 can support the establishment of a direct (e.g., wired) or wireless communication channel between electronic device 801 and external electronic devices (e.g., external electronic device 802, external electronic device 804, or server 808), and can support communication via the established communication channel. Communication module 890 may include one or more communication processors that can operate independently of processor 820 (e.g., AP), and can support direct (e.g., wired) or wireless communication. Communication module 890 may include wireless communication module 892 (e.g., cellular communication module, short-range wireless communication module, or Global Navigation Satellite System (GNSS) communication module) or wired communication module 894 (e.g., local area network (LAN) communication module or power line communication (PLC) module). A corresponding one of these communication modules can communicate via a first network 898 (e.g., a short-range communication network, such as BLUETOOTH). @ The communication module 892 communicates with external electronic devices via a second network 899 (e.g., a remote communication network, such as a cellular network, the Internet, or a computer network (e.g., a LAN or a wide area network (WAN)) or Wi-Fi Direct (or Infrared Data Association (IrDA) standard). These different types of communication modules can be implemented as a single component (e.g., a single IC) or as multiple components separate from each other (e.g., multiple ICs). The wireless communication module 892 can use user information (e.g., International Mobile Subscriber Identity (IMSI)) stored in the user identification module 896 to identify and authenticate electronic devices 801 in the communication network (such as the first network 898 or the second network 899).

[0109] Antenna module 897 can transmit or receive signals or power to or from the outside of electronic device 801 (e.g., external electronic device). Antenna module 897 may include one or more antennas. Communication module 890 (e.g., wireless communication module 892) can select at least one of one or more antennas suitable for a communication scheme used in a communication network (such as a first network 898 or a second network 899). Signals or power can then be transmitted or received between communication module 890 and external electronic device (e.g., external electronic device 802 and / or external electronic device 804) via the selected at least one antenna.

[0110] Commands or data can be sent or received between electronic device 801 and external electronic device 804 via server 808 connected to a second network 899. Each of external electronic devices 802 and 804 can be a device of the same or different type as electronic device 801. All or some of the operations to be performed at electronic device 801 can be performed at one or more of external electronic devices 802, 804, and server 808. For example, if electronic device 801 is required to automatically perform a function or service or in response to a request from a user or another device, instead of performing the function or service, or in addition to performing the function or service, electronic device 801 may request one or more external electronic devices to perform at least a portion of the function or service. The one or more external electronic devices receiving the request may perform at least a portion of the requested function or service or additional functions or services related to the request, and transmit the result of the performance to electronic device 801. Electronic device 801 may provide the result as at least part of a response to the request, whether or not the result is further processed. For this purpose, cloud computing, distributed computing, or client-server computing technologies may be used, for example.

[0111] Figure 9 This is a flowchart depicting example operations of a method 900 for predicting material properties using an AI-based material subgraph model and / or enhanced methods, according to some embodiments of this disclosure. For example, Figure 9 The illustrations depict various operations in a method 900 for predicting material properties using material subgraph models and / or enhancements, according to some embodiments. Although Figure 9 Various operations in a method according to some embodiments are illustrated, but the embodiments of this disclosure are not limited thereto, and the method may include additional or fewer operations without departing from the spirit and scope of the embodiments of this disclosure.

[0112] refer to Figure 9 Method 900 may include one or more of the following operations. Processor 212 (see...) Figure 2 Processor 212 can generate subgraphs, each of which may include the molecular substructure of the material (operation 905). Processor 212 can apply an AI-based model to the subgraphs to generate material property predictions based on the molecular substructure of the material (operation 910). Processor 212 can determine, based on the material property predictions, functions related to the material used in the production of the device (operation 915). Functions may be related to the design, manufacture and / or synthesis, testing and / or verification, and / or utilization of the material (e.g., during the manufacture of an OLED device). Processor 212 can send signals to components (e.g., production line 108, see...) Figure 1 Based on this determination, the components are controlled to perform functions related to the materials used in the production of the device (operation 920).

[0113] For example, according to some embodiments, sending signals may include sending control signals to a component operating as part of a production line, which may (e.g., in an automated manner without human intervention) physically retrieve, process, and / or manufacture materials for use as part of the device's manufacturing process. Additionally, according to some embodiments, sending signals may include sending signals to a computer system that includes a display device operating as part of a manufacturing component within a production line or production facility to display determined results thereon.

[0114] Accordingly, aspects of some embodiments of this disclosure can provide systems and / or functionalities related to AI-based material subgraph models, including: decomposition of the molecular structure of materials, generation of subgraph graphs, and implementation of subgraph modeling in a manner that allows AI-based models to be extended to large and / or complex molecules and enhances the prediction of material properties. Therefore, the disclosed embodiments can improve the overall performance of display-related products by utilizing materials deemed most suitable and / or most effective.

[0115] Embodiments of the subject matter and operation described in this specification may be implemented in digital electronic circuits or in computer software, firmware, or hardware, including the structures disclosed in this specification and their structural equivalents, or in a combination of one or more of them. Embodiments of the subject matter described in this specification may be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on a computer storage medium for execution by or control of the operation of a data processing device. Alternatively or additionally, the program instructions may be encoded on artificially generated propagation signals (e.g., machine-generated electrical, optical, or electromagnetic signals) generated to encode information for transmission to a suitable receiver device for execution by the data processing device. The computer storage medium may be, or may be included in, a computer-readable storage device, a computer-readable storage substrate, a random or serial access memory array or apparatus, or a combination thereof. Moreover, although the computer storage medium is not a propagation signal, it may be a source or destination of computer program instructions encoded in artificially generated propagation signals. The computer storage medium may also be, or may be included in, one or more separate physical components or media (e.g., multiple CDs, disks, or other storage devices). Furthermore, the operations described in this specification can be implemented as operations performed by a data processing device on data stored on one or more computer-readable storage devices or received from other sources.

[0116] While this specification may contain numerous specific implementation details, these details should not be construed as limiting the scope of any claimed subject matter, but rather as descriptions of features specific to particular embodiments. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments. Furthermore, although features may be described above as functioning in certain combinations, and even initially claimed in this way, in some cases one or more features from a claimed combination may be removed from the combination, and the claimed combination may involve sub-combinations or variations thereof.

[0117] Similarly, although operations are depicted in a specific order in the accompanying drawings, this should not be construed as requiring these operations to be performed in the specific order shown or sequentially, or requiring the execution of all illustrated operations to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. Furthermore, the separation of various system components in the above embodiments should not be construed as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0118] Therefore, specific embodiments of the subject matter have been described herein. Other embodiments are within the scope of the claims. In some cases, the actions set forth in the claims can be performed in a different order and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific order or sequence shown to achieve the desired result. In some implementations, multitasking and parallel processing may be advantageous.

[0119] As those skilled in the art will recognize, the innovative concepts described herein can be modified and varied across a wide range of applications. Accordingly, the scope of the claimed subject matter should not be limited to any specific exemplary teachings discussed above, but is defined by the claims.

Claims

1. A method for a material subgraph model, comprising: The processor generates subgraphs, each of which includes the molecular substructure of the material; The processor applies an AI-based model to the subgraph to generate material property predictions based on the molecular substructure of the material. The processor determines, based on the predicted material properties, the functions associated with the material used in the production of the target device; as well as The processor sends signals to the components to control the components to perform the functions related to the materials produced for the target device.

2. The method according to claim 1, further comprising decomposing the molecular structure of the material into the molecular substructures.

3. The method according to claim 2, wherein, The decomposition includes breaking chemical bonds that have retrosynthetic significance.

4. The method of claim 1, further comprising generating an embedding of the subgraph based on the subgraph.

5. The method according to claim 4, wherein, The embedding of the subgraph is generated by processing the subgraph through a graph neural network.

6. The method of claim 4, further comprising generating a subgraph graph based on the embedding of the subgraph.

7. The method according to claim 6, wherein, The subgraph includes nodes and edges.

8. The method according to claim 7, wherein, Each node in the node corresponds to a subgraph in the subgraph, and the value for each node in the node corresponds to an embedding in the embedding of the subgraph.

9. The method according to claim 8, wherein, Each edge in the graph represents a relationship between the subgraphs.

10. The method of claim 9, further comprising generating an updated node embedding based on the subgraph graph.

11. The method according to claim 10, wherein, Generating the updated node embeddings involves processing the subgraph graph via a graph neural network.

12. The method according to claim 9, wherein, The AI-based model analyzes the subgraph and models the relationships between the subgraphs.

13. The method according to claim 12, wherein, The target device includes an organic light-emitting diode (OLED) display device.

14. The method according to any one of claims 1 to 13, wherein, The material property prediction is based on one or more of the following: the physical properties of the material, the chemical properties of the material, the mechanical properties of the material, and the optical properties of the material.

15. An apparatus for a material subgraph model, comprising: One or more processors, the one or more processors being configured to execute: Generate subgraphs, each of which includes the molecular substructure of the material; An AI-based model is applied to the subgraph to generate material property predictions based on the molecular substructure of the material; Based on the predicted material properties, determine the functions associated with the material used in the production of the target device; as well as Signals are sent to the components to control them to perform the functions related to the materials produced for the target apparatus.

16. The apparatus according to claim 15, wherein, The one or more processors are further configured to perform: decomposing the molecular structure of the material into the molecular substructures.

17. The apparatus according to claim 16, wherein, The one or more processors are further configured to perform: generating an embedding of the subgraph based on the subgraph.

18. The apparatus according to claim 17, wherein, The one or more processors are further configured to perform: generating a subgraph graph based on the embedding of the subgraph.

19. The apparatus according to claim 18, wherein, The one or more processors are further configured to perform: analyzing the subgraph graph using the artificial intelligence-based model and modeling the relationships between the subgraphs.

20. A system for a material subgraph model, comprising: Processing circuitry; as well as A memory storing instructions that, when executed by the processing circuit, cause the processing circuit to perform the following: Generate subgraphs, each of which includes the molecular substructure of the material; An AI-based model is applied to the subgraph to generate material property predictions based on the molecular substructure of the material; Based on the predicted material properties, determine the functions associated with the material used in the production of the target device; as well as Signals are sent to the components to control them to perform the functions related to the materials produced for the target apparatus.

Images