Quaternary logic driven polymorphic structure coding method and application thereof

Abstract

The invention discloses a quaternary logic driven polymorphic structure coding method and application thereof. The technical problem to be solved is to express state, direction, hierarchy and identification integration of a nested structure, a direction path and a graph structure of data. The method comprises the following steps of mapping, identifying, encoding and forming structural unit encoding, and is applied to structural identity card encoding, graph neural network GNN structural path input encoding, storage and calculation integrated chip path encoding instructions, data authority and / or experience sharing structural identification or chip circuit encoding. In the field of artificial intelligence underlying data coding and processing, structural unit coding, direction, hierarchy and unique identification are adopted, unified coding and reversible analysis of a direction path, a nested structure and a graph topological relation of data are achieved, and the method can be widely applied to the fields of artificial intelligence, graph neural networks, chip paths, identity systems and multi-hierarchy structure modeling.

Description

Technical Field

[0001] The present invention relates to an artificial intelligence data processing method and its application, and in particular to an artificial intelligence underlying data encoding method and its application. Background Art

[0002] The mainstream data structures and AI coding underlying existing AI technologies are based on linear binary logic methods of 0s and 1s, making it difficult to efficiently express complex relationships such as nested structures, directional paths, and graph structures after processing the underlying raw data. While traditional Boolean algebra and linear ID systems are simple, they suffer from shortcomings such as a lack of path-based structural representation, hierarchical and directional labels, inability to physically map or visually read, and a lack of inherent nested logic and self-recovery structures. Summary of the Invention

[0003] The purpose of the present invention is to provide a quaternary logic driven polymorphic structure encoding method and its application, and the technical problem to be solved is to express the state, direction, hierarchy and identification of the nested structure, directional path and graph structure of data in an integrated manner.

[0004] The present invention adopts the following technical solution, a quaternary logic driven polymorphic structure encoding method, comprising the following steps:

[0005] 1. Mapping

[0006] Mapping the data of the limited original data set that can be mapped to the two-dimensional coordinate system to the nodes in the two-dimensional rectangular coordinate system;

[0007] 2. Logo

[0008] The state of the node is obtained and identified by the parity combination of the x and y values ​​of the node in the two-dimensional rectangular coordinate system. The direction, level and number of the node are identified respectively;

[0009] The direction of the node is right, up, down, left, end point or starting point;

[0010] The hierarchy means that one data or data set contains another data or data set, and one data or data set and the contained data or data set form a nested hierarchical structure;

[0011] The number identification expression node is uniquely identified in the overall structure of the data set;

[0012] 3. Coding

[0013] Encode the identifiers of the state, direction, level, and number respectively to obtain encoding units of the state, direction, level, and number;

[0014] 4. Structural unit coding

[0015] The state, direction, level and number coding units are combined in the order of state coding unit, direction coding unit, level coding unit and number coding unit to form the structural unit coding of the node.

[0016] The method of the present invention has four states: state 1, indicating that the x and y coordinate values ​​of the node are even, even; state 2, indicating that the x and y coordinate values ​​of the node are even, odd; state 3, indicating that the x and y coordinate values ​​of the node are odd, even; state S4, indicating that the x and y coordinate values ​​of the node are odd, odd.

[0017] The state coding unit of the method of the present invention is represented by four digits 1 and 0, the state 1 code is 0001, the state 2 code is 0010, the state 3 code is 0100, and the state 4 code is 1000.

[0018] The direction coding unit of the method of the present invention is represented by four digits 1 and 0, the right direction code is 0001, the upward direction code is 0010, the downward direction code is 0100, the left direction code is 1000, the inward direction code is 0000, and the outward direction code is 1111.

[0019] The hierarchical coding unit of the method of the present invention is represented by four-digit numbers 1 and 0, the first to fourth layers are represented by four-digit numbers, and the fifth layer and above are represented by at least eight-bit sparse coding.

[0020] The method number encoding unit of the present invention is selected by one of the following methods:

[0021] (1) Coding structure: the number field adopts a 4-bit, 8-bit, 12-bit, or 16-bit sparse activation structure, and the bit width is expanded according to the size of the data set and the required coding information dimension;

[0022] (2) Sparse activation, activating 1 to 4 bits with semantic meaning in the encoding, and keeping the rest at 0;

[0023] (3) Bit segment grouping, where the numbered bit pattern is divided into at least two functional segments;

[0024] (4) Symmetrical coding, using mirror image comparison, positive and negative states;

[0025] (5) Semantic parsing: Each set of activation bits can be reverse parsed through semantic mapping.

[0026] The method number identification of the present invention also expresses the multiple semantic mappings of the node in the overall structure of the data set.

[0027] The method of the present invention has multiple semantic mappings with a semantic parsing control table, and the semantic parsing control table is set and determined during encoding.

[0028] The method of the present invention switches multiple semantic mappings by using one of the following switching methods:

[0029] (1) Encoding built-in control bit type switching

[0030] Embed a 4-bit structure control code in the high bit of the number field or the reserved field as a switching identifier;

[0031] (2) Context Binding

[0032] Bind the semantic control table according to the scenario when calling the external system;

[0033] (3) External protocol or structure identification trigger

[0034] Append control instruction flags to structure data flow or task descriptions.

[0035] An application of a quaternary logic-driven polymorphic structural encoding method is applied to structural ID card encoding, graph neural network (GNN) structural path input encoding, storage and computing integrated chip path encoding instructions, data permission and / or experience sharing structural identification or chip circuit encoding.

[0036] Compared with the existing technology, in the field of artificial intelligence underlying data coding and processing, the present invention adopts structural unit coding, direction, hierarchy and unique identification to realize the unified coding and reversible analysis of the data's directional path, nested structure, and graph topological relationship, and constructs a highly compatible multi-purpose coding system that can be widely used in artificial intelligence, graph neural networks, chip paths, identity systems, and multi-level structure modeling, and shows significant advantages in expressiveness, energy consumption control, path tracing and tensor modeling. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is the four-element state logic coordinate diagram of the present invention.

[0038] Figure 2 This is a composition diagram of the 4HPU structural unit of the present invention.

[0039] Figure 3 It is a schematic diagram of the standard expression of the structural unit of the present invention.

[0040] Figure 4 This is the visual chip lighting logic diagram of the present invention.

[0041] Figure 5 It is the structural path tensor diagram of the present invention.

[0042] Figure 6 It is a schematic diagram of the relationship between hierarchical fields and hierarchical nesting of the present invention.

[0043] Figure 7 This is a schematic diagram of the organizational structure coding of a small company according to an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0045] The present invention's quaternary logic-driven polymorphic structure encoding method (method) is used for encoding raw data at the bottom layer of artificial intelligence. It uses structural unit 4HPU encoding, which is composed of encoding units. The encoding units are: encoding units for a state identifier (state) S, a direction identifier (direction) D, a level identifier (level) L, and a number identifier (unique identifier) ​​ID. The S encoding unit consists of two digits (two-bit type), and the D, L, and ID encoding units consist of four digits (four-bit type). The encoding units S, D, L, and ID are combined in this order to form a structural unit code of S+D+L+ID, which is used to express each data in the bottom layer of artificial intelligence data set.

[0046] The method of the present invention establishes the field codes of the data state S, direction D, level L and number ID respectively, and then arranges them in sequence to form a structural unit code. Figure 2 As shown, the data in the limited original dataset that can be mapped to a two-dimensional coordinate system is first mapped to nodes (structural elements) in a two-dimensional rectangular coordinate system. The node state S is obtained and identified by combining the parity of the x and y values ​​of the node in the two-dimensional rectangular coordinate system. Then, the direction, level, and number of the node are identified according to the node direction D, level L, and number ID. The state S, direction D, level L, and number ID are further encoded to obtain the encoding unit of S, D, L, and ID. Finally, the encoding units S, D, L, and ID are sequentially combined to form the structural unit encoding of the node.

[0047] Direction, hierarchy and unique identification realize the unified encoding and reversible analysis of data's directional path, nested structure and graph topological relationship.

[0048] The method of the present invention is based on a two-dimensional coordinate system and a four-field encoding method, called a 4HPU (4-state Holographic Processing Unit) structural encoding unit, which is expressed as follows: the structural unit of each node contains: one of the four states determined by the parity of the x and y values ​​+ a direction identifier + a hierarchical nesting identifier + a number field, which can express any complex path, nesting, branching and dynamic relationship of data or data sets, and can be reconstructed (restored to the original value of the data) without relying on other protocols or interpreters. Figure 3 As shown, the encoded data identifier is represented as: 4HPU (S, D, L, ID).

[0049] like Figure 5As shown, the original data in the dataset is mapped to a two-dimensional rectangular coordinate system. Each original data is a node. The encoding method of the node S, D, L, and ID is:

[0050] 1. Status field S encoding

[0051] The state field determines the state of a node by the parity combination of the x and y values ​​in the two-dimensional coordinates (x, y). There are four states (four-color state, four-element state): S1, S2, S3 and S4. The state code is used to distinguish the state category of the node when classifying the basic position of the node in the mapping data structure. The state code uses the four-digit 1 and 04HPU encoding. Figure 1 As shown, the four states are:

[0052] State S1 indicates that the x and y coordinate values ​​of a node are even, even, and the 4HPU code is 0001.

[0053] State S2 indicates that the x and y coordinate values ​​of this node are even and odd, and the 4HPU code is 0010.

[0054] State S3 indicates that the x and y coordinate values ​​of this node are odd or even, and the 4HPU code is 0100.

[0055] State S4 indicates that the x and y coordinate values ​​of this node are odd and odd, and the 4HPU code is 1000.

[0056] 2. Direction field D encoding

[0057] The direction field is used to describe the directional relationship between nodes in the structural path. There are six identifiers for expressing directions. The direction code is used when establishing a path connection from the current node to the adjacent node. The direction code is represented by four digits 1 and 0. The structural path refers to a sequence of directed connections between nodes in the network formed between data nodes based on spatial position relationships or flow direction relationships. The structural path is used to characterize the directional direction (flow), hierarchical connection or logical order between nodes. Specifically:

[0058] The symbol for the node to the right is: →, the direction code is 0001, and the right direction code is used when establishing a path connection from the current node to the right adjacent node.

[0059] The symbol for the node upward is: ↑, and the direction code is 0010. The upward direction code is used when establishing a path connection from the current node to the adjacent node.

[0060] The downward symbol of the node is: ↓, and the direction code is 0100. The downward direction code is used when establishing a path connection from the current node to the adjacent node.

[0061] The symbol for the node to the left is: ←, the direction code is 1000, and the left direction code is used when establishing a path connection from the current node to the left adjacent node.

[0062] The symbols for nodes pointing inward are: The direction code is 0000. The 0000 direction code is used when establishing the end point of the structural path, that is, connecting to the current node to end the data node flow, serving as the end node of the path or the internal convergence node.

[0063] The outward symbol of the node is: ⊙, and the direction code is 1111. The 1111 direction code is used when establishing the starting point of the structure path, that is, initiating a connection or data node flow from the current node as the starting node of the path.

[0064] The direction field is represented by a four-bit sparse code, and together with the status field S, the level field L and the number field ID, it forms a complete 4HPU structure coding unit, which is used to represent node connections, path tracing and dynamic structure expansion.

[0065] Sparse structure refers to a locally focused, well-defined pattern of activation bits within a uniform length, such as a 4-bit or 8-bit coded pattern. It is used to express structural direction, path characteristics, or symbol state. In the coded pattern, positions displaying 1 are active, while positions displaying 0 are inactive.

[0066] The determination of sparse structures in the method of the present invention is based on the following three features: 1. semantic activation correspondence, the activation bit must correspond to a predefined structural meaning, such as a direction mapping table; 2. symmetric structural regularity, the activation pattern should be able to maintain the positive and negative symmetric relationship of the encoding, such as positive 0 is encoded as 0000, negative 0 is encoded as 1111, the positive direction is encoded as 0001, and the negative direction is encoded as 1110, that is, each positive state (such as positive S class: S1, S2, S3, S4) has a mapped corresponding negative state (such as negative T class: T1, T2, T3, T4); 3. locality of the activation area, the activation bit presents concentrated activation or mirror activation in the logical structure to avoid global diffuse expression. For example, the code 0001 represents rightward (→), which conforms to the unit-activated sparse structure. The code 1110 represents a negative leftward (←) path. Although it activates 3 bits, it is a symmetrical expression of the negative state and is also considered to be structurally sparse. The code 1111 (negative 0) is a special termination symbol and belongs to a closed state bit type, which is also within the scope of structurally sparse expression.

[0067] A sparse structure is a coding bit pattern of uniform length. Although the number of activated bits may vary from 1 to 4, its activation pattern must possess structural characteristics, and the mapping, path direction, or state identifiers must form a one-to-one correspondence. The core of a sparse structure lies in: the activation bits are non-spread and semantically focused; the structure they represent has clear meaning, and the path identifiers are unique or resolvable, which can be used for positive and negative paths, start and end states, and loop nesting. For example: 0001 and 0010 are basic sparse path direction codes; 1110 and 1101 are symmetrical reverse expressions, still possessing sparse structural characteristics; encoding 1111 (negative 0) activates all bits and serves as a structural path termination mark or "closer." Although 4 bits are activated, the structure is highly focused and still belongs to a sparse structure. Therefore, sparse structures differ from the traditional computer system definition of "only one bit is 1" and emphasize the centralization of expression logic, the consistency of path direction, and the completeness of the structure.

[0068] Sparse structure coding has the following advantages: the node activation path is clear and physically easy to map into controllable signals (such as lighting up lights and connecting lines); the expression is sparse and redundancy is avoided, making it suitable for low-power and edge device scenarios; it is highly consistent with the direct translation and direct reading characteristics of the 4HPU architecture and can be directly transmitted and used without decoding.

[0069] The jump of each node in the structural path is indicated by a direction code, forming a change in the structural topological distribution.

[0070] A jump refers to a node moving from its current position in a specified direction or connecting to another node's position according to the direction field D, forming a directed connection between nodes and constructing a topological distribution change of the entire data structure. Jumps not only describe the connection behavior between nodes, but also define the directionality of data or state flow. In a multi-layer nested structure, jumps can span different levels. Jumps are the basic behavior for constructing changes in the topological distribution of structures, representing spatial migration, logical flow, hierarchical recursion, and entry into subspaces between nodes. Each jump action is indicated by a corresponding direction code (D field) to ensure the consistency of the data structure in application scenarios such as path construction, trajectory tracking, and dynamic management.

[0071] 3. Level field L encoding

[0072] The level field indicates that a data set or dataset contains another data set or dataset. This indicates that a data set or dataset and the contained data set form a nested hierarchical structure, which is called the contained data set or dataset being nested within the contained data set or dataset. This field indicates the depth of the nesting of the nodes. The level field, L, is represented using a sparse four-bit code.

[0073] Sparse coding refers to activating a small number of one or two significant bits within a uniform-length coding bit pattern, while the remaining bits are displayed as zeros and remain off. Alternatively, all bits can be activated in specific circumstances (e.g., the special end flag 1111), or no bits can be activated (e.g., the default empty layer flag 0000). This achieves code sparsity, structural focus, and efficient expression. Valid bits in a coding bit pattern are those that are mapped to actual structural functions (e.g., direction, level, or state) within the structural path. For example, bit 1 in 0001 indicates "path to the right," and 1 is a valid bit. In 0000, no bits are activated, but in the level field, it can represent "default level," making the structure valid overall. In 1111, all bits are activated, marking "structural termination or closed path," a legal sparse structure with a negative state. Valid bits are determined primarily based on two mechanisms: a position mapping table, which predefines the direction, state, or path meaning for each bit position; and contextual semantic binding, which interprets the validity of an activated bit's function based on the field type (e.g., direction D, level L). Structural information is expressed through the position of the representation rather than the overall bit value. The overall data structure exhibits a sparse structure characteristic, that is, in a large-scale node distribution, only a small number of node paths are dynamically represented according to actual needs, and the others remain inactive, thereby achieving storage resource optimization, reduced computational complexity, and distributed elastic expansion. For example:

[0074] In the 4-bit fixed encoding space, there are theoretically 16 combinations.

[0075] The coding bit patterns used in actual high frequency use only include the basic four states: 0001, 0010, 0100 and 1000, which are used for unidirectional paths, hierarchical identification or state expression respectively;

[0076] Positive and negative zero states: 0000 indicates the default or empty state, and 1111 indicates the closed or negative zero state;

[0077] Negative symmetric extended states: 1110, 1101, 1011, and 0111, used to represent the reverse semantics of direction, state, or path.

[0078] In addition to the above main types, such as: 0011, 0101, 1100, 1010, the combination exists in theory, but in the standard structure

[0079] It is almost not used in the structural path, hierarchical expression and control logic, avoiding redundant representation, which shows that the encoding method of the present invention has obvious structural sparse distribution characteristics.

[0080] like Figure 6 As shown in Figure 1, in hierarchical coding, a four-bit sparse structure format is used to express the nested relationship of each layer, from the outer layer to the inner layer: L1, L2, L3, L4, L5. The specific definitions are shown in Table 1.

[0081] Table 1 Definition of hierarchical coding

[0082]

[0083] The first layer is identified as L1, and the level code is 0001. When the node is nested and divided into levels, the level code is used to indicate that it is in the first level of the overall structure, that is, the outermost or upper position. The level code is represented by four digits.

[0084] The second layer is identified as L2, and the level code is 0010, indicating the second layer, which is nested in the first layer.

[0085] The third layer is identified as L3, and the layer code is 0100, indicating the third layer, which is nested in the second layer.

[0086] The fourth layer is identified as L4, and the layer code is 1000, indicating the fourth layer, which is nested in the third layer.

[0087] The fifth layer and above use extended bit type representation:

[0088] The fifth layer is identified as L5, and the level code is 0001 0000. The 4HPU sparse structure code corresponding to the natural number 5 represents the fifth layer, which is nested in the fourth layer.

[0089] The sixth layer is identified as L6 and the layer code is 0010 0000.

[0090] And so on.

[0091] At least eight-bit sparse coding is used above the fifth layer.

[0092] Nesting levels can be expanded to any valid number within a preset upper limit. The specific upper limit can be set according to the application scenario requirements, ranging from a single layer to hundreds of layers or even higher. The corresponding level field is represented in the 4HPU sparse coding format. That is, each nested structure is encoded with a 4-bit bit type to identify its level status, enabling hierarchical identification and management in the structure path.

[0093] 4. Number field ID encoding

[0094] The number field is used to express the unique identification or multiple semantic mapping of the node in the overall structure of the dataset.

[0095] Unique identification uses a sparse bit representation of the ID field to assign a unique and non-repeatable identity code to each node within the dataset. The ID remains constant despite node spatial relocation, path reconfiguration, and level shifts, ensuring fast node retrieval, independent indexing, and path traceability. The ID field also supports multiple semantic mappings, encoding bit types with attributes such as path history, hierarchical attribution, time period classification, category, and tag.

[0096] Multiple semantic mappings mean that the ID field not only identifies the node but can also be mapped to different semantic types based on parsing rules, such as natural number numbers, status classifications, path order, hierarchy, time and batch, or special markers (such as abnormal nodes, closed states, or entry points). This enhances the expressiveness and application flexibility of the coding system. This multi-mapping feature allows the same set of ID codes to be parsed differently in different application scenarios and contexts, greatly improving the expressiveness and reuse efficiency of the coding system.

[0097] The parsing rules are a pre-set semantic parsing control table, for example, see Table 2.

[0098] Table 2 Semantic parsing control table

[0099]

[0100] The high, middle, and low bits in Table 2 refer to the three groups of 4-bit encoding arrangements, where the left group is the high bit, the middle group is the middle bit, and the right group is the low bit.

[0101] The semantic parsing control table is usually set and determined during encoding. It can be switched and identified during encoding or different application scenarios. The switching method and identification include the following three categories. Choose one:

[0102] 1. Encoding built-in control bit switching

[0103] The control bit type is used as the decoding judgment basis when the structure is entered, and the corresponding semantic analysis control table is bound.

[0104] A 4-bit structure control code is embedded in the high bit of the ID field or the reserved field as a switch identifier, see Table 3.

[0105] Table 3 Switching flag

[0106] Control bit type Semantic meaning 0001 Structural path semantic pattern 0010 Permission and role label semantic model 0100 Semantic patterns of timing and state signals 1000 Physical mapping and chip path control mode 0000 Default general mode (no need to switch)

[0107] 2. Context Binding

[0108] It does not embed coded control bits, but instead binds semantic control tables based on the scenario when an external system calls it. This makes it suitable for graph neural networks and multi-module integration environments. For example, a GNN module automatically binds to a "structural path mode," and an AI inference system automatically loads a "state mapping rule set."

[0109] 3. External protocol or structure identification trigger

[0110] Add control instruction flags in the structured data flow or task description, such as indicating the "activation mode code" through an external field, and dynamically load the corresponding semantic switch.

[0111] In this way, flexible expansion and multi-purpose scenario applications can be carried out under unified coding rules.

[0112] Multiple semantic mappings In 4HPU encoding, the same number field ID can be mapped to multiple semantic roles with functional distinctions through context binding and combination with the semantic table (see Table 4), rather than relying on the numerical number obtained by accumulating bit weights.

[0113] Table 4 Semantic table

[0114]

[0115]

[0116] The ID field is represented using a sparse bit type. Within a uniform length (e.g., 4, 8, or 16-bit) encoding bit type, only a few structurally significant bits are activated (set to 1), while the remaining bits remain disabled (set to 0). The node's unique identifier is formed by the combination of these bits. This bit type representation can be based on node spatial position, path order, hierarchical depth, or decentralized identity identification (DID) mapping, ensuring efficient, low-conflict, and scalable identity management. Specifically, within a fixed-length encoding bit type (e.g., 4, 8, or 16-bit), only a very small number of specific bits are set to 1. The activation of these bits is determined by: the node's logical spatial position (e.g., row and column in a zigzag numbering scheme), its path role (e.g., entry, turn, or endpoint), its category, functional type, or identity within the system, or an external identifier derived from a binding (e.g., the low-order bits of a DID hash). A unique, reproducible combination of these activated bits is mapped to form the node's ID field. The remaining bits are set to 0, and the combination of these bits serves as the node's unique identifier.

[0117] For example, the ID field of a node is encoded as: 0001 0010. In this 8-bit sparse structure bit type: the upper 4 bits: 0001 indicate that the node has the "entry structure" semantics, and the lower 4 bits: 0010 indicate that the node is in the "2nd layer, second from the left position" in the structural hierarchy. The node has a uniquely locatable combined identity in the spatial structure.

[0118] If numbered from right to left by bit type, this code activates bit 2 (weight 2), the second position in the lower layer, and bit 5 (weight 16), representing high-level structure type B. The resulting structure is "Type B entry node x second path position." Because the positional combination of activated bits has fixed semantic meaning and cannot be remapped within the system, this code constitutes the globally unique identifier of a node in the network. It can be used for unique path identification, graph node tracking, access mapping for integrated storage and computing chips, distributed AI memory indexing, and network structure traceability.

[0119] The ID field is encoded using one of the following methods:

[0120] 1. Coding structure: The ID field usually adopts a 4-bit, 8-bit, 12-bit, or 16-bit sparse activation structure, and the bit width is expanded according to the size of the dataset and the required encoding information dimension.

[0121] 2. Sparse activation method: only a few (1 to 4 bits) with semantic meaning are activated in the encoding, and the rest remain 0 to meet the needs of information sparsity, high-speed decoding and physical mapping.

[0122] 3. Bit segment grouping method, the ID bit type can be divided into multiple functional segments, which are listed in Table 5. In the encoding method of the number field ID, bit segment grouping is preferred.

[0123] Table 5 ID encoding bit segment grouping

[0124] Bit field Functional Examples High 4 bits Node classification codes, such as type and semantic labels. Middle 4 Path flow characteristics, such as inbound and outbound directions, and status identification. Low 4 to 8 bits A unique code for a node, such as a serial number or hash bits.

[0125] 4. Use symmetric encoding. To satisfy the mapping of positive and negative states, ID encoding allows certain structures to use symmetric bit types, such as 0001 and 1110 for path tracing, mirror comparison, and positive and negative states.

[0126] 5. Semantic parsing method: each set of activation bits can be reverse parsed into node meaning through semantic mapping, such as:

[0127] 00010010 can be parsed as: category B, ingress direction, number 18.

[0128] 01000001 can be represented as: terminal node, internal jump, number 65.

[0129] The code 00010010 can be interpreted as the natural number 7. In the 4HPU path structure, the sparse code of 18 is 01000100 (i.e. S3→S3); 00010010 belongs to the sparse activation bit type, according to the bit weight (2nd bit + 5th bit)

[0130] It is 2+16=18.

[0131] In this regard, whether it is parsed as a natural number depends on the semantic parsing control table bound to the context.

[0132] 6. Conflict detection method: When allocating IDs, a mapping hash table or sparse bitmap pool should be set to prevent the same code from being reused in the same data set and ensure the global uniqueness of the coding structure.

[0133] When a node's 4HPU code is (S2, D→, L1, ID = 00010010), it means: the state field S takes the value S2 (even-odd), which is encoded as 0010; the direction field D indicates a right connection, which is encoded as 0001; the level field L is at the first level (L1), which is encoded as 0001; and the code field ID has a value of 00010010, which uniquely identifies the node within the overall structure. The 4HPU code is: 0010 0001 0001 0001 00010010.

[0134] The ID coding method of the present invention can be used in low-power engineering practices of hardware coding, chip burning, and photoelectric coding identification, and is particularly suitable for IoT device identification and identity registration and tracking tasks in large-scale node networks.

[0135] like Figure 7 As shown in the figure, a small organizational structure dataset of a company is used as an example, which contains five nodes:

[0136] General Manager, set as node A,

[0137] Department manager B is set as node B.

[0138] Department manager C is set as node C,

[0139] Ordinary employee D is set as node D,

[0140] Ordinary employee E is set as node E,

[0141] Map the above nodes to a two-dimensional rectangular coordinate system. The original position of each node is determined by the parity of the state field S. There is a superior-subordinate command relationship between the nodes, which is defined as the direction field D of the path. The company forms a (nested) hierarchy field L from the top to the bottom, and each node is assigned a unique number field ID. According to the above settings,

[0142] Generate a quaternion state based on the parity combination of x and y of the node in the two-dimensional coordinate system:

[0143] Even-even: 0001 means state 1S1,

[0144] Even-odd: 0010 means state 2S2,

[0145] Parity: 0100 means state 3S3,

[0146] Qiqi: 1000 means state 4S4.

[0147] The relationship between the nodes is:

[0148] General Manager A issues instructions to department managers B and C. The direction of the path is from A to B and C.

[0149] Department manager B issues instructions to ordinary employee D. The direction of the path is from B to D.

[0150] Department manager C issues instructions to ordinary employee E, and the direction of the path is from C to E.

[0151] The level field L is defined as three levels:

[0152] The general manager is at the first level L1, with the code 0001.

[0153] The department manager is the second level L2, coded as 0010.

[0154] Ordinary employees are at the third level L3, coded as 0100.

[0155] Number field ID, each node is assigned an independent sparse bit type ID: IDA, IDB, IDC, IDD, IDE.

[0156] The coding of this embodiment is shown in Table 6.

[0157] Table 6 Company's three-tier organizational structure code

[0158]

[0159] In the example of a company's organizational structure, the general manager node A is connected to the department manager nodes B and C via a direction; department manager B is connected to ordinary employee D via a direction; and department manager C is connected to ordinary employee E via a direction. This forms a directed structural path network that clearly expresses the direction of command flow within the organization.

[0160] The method, encoding steps and technical effects of the present invention are compared in Table 7.

[0161] Table 7 Comparison of coding steps and technical effects

[0162]

[0163] The method of the present invention can be further expanded:

[0164] 1. All codes can be combined into a sequence of structural paths

[0165] In the present invention, the structural unit codes of all nodes (i.e., the state field S, the direction field D, the level field L, and the number field ID) can be combined to form a structural path sequence. The specific method is: according to the connection relationship between the nodes, the 4HPU coding units of the nodes are arranged in sequence according to the indication order of the direction field D to form a directed path sequence. If the node jumps from state S1 (even-even) to state S3 (odd-even) via a path, and then jumps to state S4 (odd-odd), the corresponding structural path sequence is: from S1 via S3 to S4. Each jump is indicated by the direction D field, while retaining the node's own state S, level L, and number ID information to achieve complete structural expression.

[0166] 2. Multi-layer encoding state mapping, encoding can be nested hierarchically

[0167] The four-field encoding (S, D, L, ID) of the present invention can be used for multi-layer nesting, that is, multiple sub-node encodings can be embedded in a node encoding to form a recursive structure mapping. The specific process is: the outer node defines the basic S, D, L, ID; the inner node redefines its own S, D, L, ID in a nested manner; multiple layers of recursive combination form a hierarchical path structure similar to nested brackets. For example: [S1[S2[S4]]], which means: node S2 is nested under node S1, and node S4 is nested under node S2. Each layer retains the complete 4HPU field information.

[0168] 3. Express the controllable visual chip lighting mode (bit type representation)

[0169] In the method of the present invention, the sparse representation bit pattern of the node or path encoding can directly indicate the chip lighting mode, realizing the visual expression of the structure. Figure 4 As shown in the figure, the chip lighting mode is as follows: each sparse bit pattern (such as 0000, 0001, 0010, 0000) corresponds to a physical lighting position in the chip. Activating a bit pattern illuminates the physical lamp bead or signal node at the corresponding position. The node's S, D, L, and ID fields can all be mapped to a visible lighting state through bit activation. For example, the code 0001 for node S1 activates the first light position on the chip, forming a visible structural path diagram.

[0170] 4. Compressed encoding (state compression) is possible

[0171] In this invention, the state field S can be compressed to improve storage and transmission efficiency. The specific process is to remap the four states (S1, S2, S3, and S4) originally encoded in a 4-bit sparse representation to a 2-bit compact code (00, 01, 10, and 11). This achieves a compressed representation of the state information, reduces the overall code length, and improves data processing speed.

[0172] 5. Extended coding (disturbance factor and / or modulo N expansion) is possible

[0173] The present invention allows for code expansion to enhance flexibility and complexity. Specific expansion methods include: adding perturbation factors, superimposing timestamps, environmental tags, and random perturbation codes on the original code to form a dynamically changing code; and Modulo-N expansion, which expands state classification from the basic 4 states to Modulo-N states (e.g., Modulo-8 and Modulo-16), enabling state subdivision and dynamic state expression in more complex scenarios.

[0174] 6. Adapting Graph Neural Network Tensor Input, Structured Identity Model, and Path Backtracking Methods

[0175] The four-field 4HPU encoding of this invention is suitable for graph neural network (GNN) tensor input and structured identity modeling. The process is as follows: the S, D, L, and ID codes of each node are combined as a node feature vector in a high-dimensional tensor; the directional D field drives the connection relationship between nodes to form the directed edges of the graph; the overall structured identity model is constructed, which can be used for individual identification, trajectory backtracking, and anomaly detection. By tracing back along the encoded path (according to the reverse mapping rules of the D field), the node path can be reversibly restored and the flow trajectory can be recovered.

[0176] 7. The encoding path is reversible

[0177] The structural path encoding of the present invention is reversible. The process is as follows: each jump records the direction D; through the reverse mapping of direction D (e.g., correspondence, correspondence), the connection relationship between nodes can be reversed; the hierarchical L field can be nested and expanded, so that each layer of the substructure can be recursively traced back; the ID number ensures that the node identity remains unchanged during the tracing process, allowing reliable path restoration.

[0178] In the method of the present invention, the number field ID is used for sparse representation, the reversible expression of the structural path, the hierarchical nesting and the activation bit combination are used for compressible expansion rules and visual lighting control, and the activation bit controls the physical LED for visual lighting control.

[0179] The method of the present invention can be applied to multiple fields such as artificial intelligence underlying coding, structural graph modeling, chip path control, identity modeling and AI tensor input.

[0180] The method of the present invention is based on the 4HPU structural system, has good uniformity, scalability and engineering adaptability, and can be used for: structural ID card coding, AI Agent local memory coding, graph neural network GNN structural path, storage and computing integrated chip path coding, Chinese coding standard alternatives, data authority and / or experience sharing structural identification, identity modeling, path recognition, chip path control, tensor input, structural visualization system, physical steps (such as chip circuits, light paths) scenarios, see Table 8. The structural visualization system is to intuitively display the image of the 4HPU coding nodes and paths by lighting the chip, dynamically displaying the graphical interface, and activating the light path. The physical step refers to the "physical entity implementation step of chip circuit wiring and light path design", that is, directly mapping the 4HPU structural path to the connection or path relationship in the actual physical hardware.

[0181] Table 8 Application scenarios of the method of the present invention

[0182]

[0183]

[0184] The following further lists the encoding of 5 application scenarios.

[0185] Scenario 1: Structural ID Card Encoding

[0186] The first step is the original data, a city resident database, each resident record has: location coordinates (x, y), residential building number, identity unique ID.

[0187] The second step is to extract the four fields, see Table 9.

[0188] Table 9 Four fields extraction

[0189] Field content S (status field) <![CDATA[Determine S1 to S4 based on the parity of the x and y coordinates]]> D (direction field) The starting point is set to ⊙, with no flow direction L (Level field) Residential area level mark, L1, L2 ID (number field) Sparse representation coding based on resident ID mapping

[0190] The third step is to form the 4HPU code, code: 4HPU (S01, D⊙, L1, ID = 00010010).

[0191] The fourth step is to input the coded structure path into the application system composed of computer systems. Each resident node generates a structural ID card using 4HPU encoding. It can be used for: urban distribution visualization, identity verification system, and smart city governance.

[0192] Scenario 2: Graph Neural Network (GNN) Structure Path Input Encoding

[0193] The first step is the original data, a traffic network graph, where each node is a road intersection and each edge is a road.

[0194] The second step is to extract the four fields, see Table 10.

[0195] Table 10 Four fields extracted

[0196]

[0197]

[0198] In the third step, the 4HPU code is formed, 4HPU (S03, D→, L2, ID = 00100100).

[0199] In the fourth step, the structured path is input into the application system composed of computer systems. The node codes are input into the GNN as feature vectors, and the direction D establishes the directed edges of the graph. This can be used by AI models for traffic flow prediction and path optimization.

[0200] Scenario 3: Path encoding instructions for integrated storage and computing chips

[0201] The first step is to create raw data and a network of integrated storage and computing units inside the chip, with instruction flow connections required between units.

[0202] The second step is to extract the four fields and convert Table 11.

[0203] Table 11 Four fields extraction

[0204] Field content S Unit position parity determination D Data flow direction (→, ↑, etc.) L Functional level (e.g. arithmetic unit / control unit) ID Unit number sparse activation bit pattern

[0205] The third step is to form the 4HPU code, 4HPU (S04, D↓, L1, ID = 01000010).

[0206] In the fourth step, the structured path is input into the application system composed of the computer system. The internal data flow path of the chip is directly controlled by the 4HPU encoding instructions. This can be used for dynamic path reconstruction and low-latency data flow.

[0207] Scenario 4: Data permissions and / or experience sharing structure identification

[0208] The first step is the original data, an online education platform, user experience sharing, including learning records and knowledge fragments.

[0209] The second step is to extract the four fields and convert Table 12.

[0210] Table 12 Four fields extraction

[0211]

[0212]

[0213] In the third step, the 4HPU code is formed, 4HPU (S02, D→, L2, ID = 00100010).

[0214] The fourth step is to input the structure path into the application system composed of the computer system. Each sharing unit is managed with 4HPU code for fast authorization, permission verification, and traceability.

[0215] Scenario 5: Chip circuit coding process

[0216] The first step is the original data, a chip logic module, circuit nodes and connection paths.

[0217] The second step is to extract the four fields, see Table 13.

[0218] Table 13 Four field extraction

[0219] Field content S Circuit node position status (odd or even) D Direction of electrical signal flow L Circuit layer (such as logic layer, storage layer) ID Node number sparse activation

[0220] In the third step, the 4HPU code is formed, 4HPU (S01, D←, L1, ID = 00011000).

[0221] In the fourth step, the structural path is input into the application system composed of the computer system, and the 4HPU code is used to guide the physical connection design. This is used to simplify wiring optimization and improve circuit stability and area utilization.

[0222] The 4HPU code segment formed by the method of the present invention can express each node or structural element as follows: state classification (S) + path direction (D) + nesting level (L) + ID path (I); as a global unique identifier of the data set, it can be reversibly restored, that is, each group of 4HPU code segments has a reversible parsing path, such as:

[0223] The code is 00010010, which can be parsed into different meanings:

[0224] In natural number parsing mode, this bit pattern activates bits 2 and 5, which can be calculated as the natural number 7;

[0225] In the state sequence parsing mode, the bit pattern can be split into two 4-bit segments 0001+0010, representing states S1 and S2 respectively, which can also be regarded as the state activation sequence in the two-layer path structure.

[0226] Switching is determined by the current application context or control bit type. The "state sequence semantic table" is bound to the structure path, the "natural number mapping rule" is bound to AI nested recognition, or a control code such as 0001 or 0010 is embedded in the high bit of the ID to indicate the current parsing mode.

[0227] The present invention supports multiple semantic parsing and dynamic switching capabilities for the same set of codes under a unified coding type through a preset control type or a system context binding mechanism.

[0228] The multi-layer structure uses the direction field D to map the path, combined with the level field L to distinguish nested relationships. Ultimately, together with the status field S and the number field ID, it forms a complete 4HPU structure encoding path. Therefore, based on the 4HPU encoding segment, hierarchical backtracking, path tracing, and node feature vector reconstruction can be performed.

[0229] In the method of the present invention, the encoding steps based on the sparse structure of the number field ID, the structural connection mapping of the direction field D, the path nesting identification of the level field L, and the position status expression of the status field S have the following five advantages, as shown in Table 14.

[0230] (1) Unified coding structure, based on the unified structural coding steps of "status field S + direction field D + level field L + number field ID", realizes the "quaternary fusion coding" of status, path, level and identity, and solves the splitting redundancy problem of traditional binary logic when expressing complex structural relationships.

[0231] (2) Nested path support: Based on the combined structure construction steps of "direction field D and level field L", it supports multi-layer nesting and context recursion of path nodes, and can express any complex path topology relationship in the graph structure, making up for the defect that the linear numbering system is difficult to express nested structures.

[0232] (3) Sparse structure visualization control. In the sparse structure generation step of the number field ID and the direction field D, the activated specific bits can be directly mapped to LED light positions, chip channels, or signal control elements, facilitating the physical visualization lighting of the structure path and low-power hardware mapping control.

[0233] (4) Compression and expansion capabilities: In the path combination and hierarchical encoding steps, any path can be encoded and compressed by combining and expanding the activation bits. At the same time, it supports multi-layer recursive expansion and adapts to diverse scenarios with structural density ranging from low (short paths) to high (deep nesting).

[0234] (5) Path coding reversibility. Based on the structural combination and sparse structure decodability characteristics of the method of the present invention, the original state of the structure and path topology can be quickly restored through local decoding algorithms, path backtracking and bitmap indexing, and offline path restoration and reconstruction analysis can be performed in smart terminals and edge computing scenarios.

[0235] Table 14 Five advantages

[0236] serial number Technical Effects The structural encoding steps derived from (1) Integration of status, direction, level and ID Four-field structure combination steps (S+D+L+ID) (2) Multi-layer nesting and path recursive expression D+L coding combination nested path steps (3) Visual hardware control lights Sparse structure coding step (ID / D activation bit) (4) Compression and expansion support Path-level recursive compression combination steps (5) Reversibility + Pathway Reconstruction Sparse structure + hierarchical traceable path parsing steps

[0237] In summary, the present invention provides a new structural expression system that is unified, efficient, scalable, and highly adaptable in the field of underlying data coding and processing in artificial intelligence.

[0238] The present invention constructs a highly compatible multi-purpose coding system through sparse structure expression, path compression mechanism and activation bit visual control. It can be widely applied to fields such as artificial intelligence, graph neural networks, chip paths, identity systems and multi-level structure modeling, and shows significant advantages in expressiveness, energy consumption control, path tracing and tensor modeling.

Claims

1. A polymorphic structure encoding method driven by quadruple logic, comprising the following steps:

1. Mapping Mapping the data of the limited original data set that can be mapped to the two-dimensional coordinate system to the nodes in the two-dimensional rectangular coordinate system; 2. Logo The state of the node is obtained and identified by the parity combination of the x and y values ​​of the node in the two-dimensional rectangular coordinate system. The direction, level and number of the node are identified respectively; The direction of the node is right, up, down, left, end point or starting point; The hierarchy means that one data or data set contains another data or data set, and one data or data set and the contained data or data set form a nested hierarchical structure; The number identification expression node is uniquely identified in the overall structure of the data set; 3. Coding Encode the identifiers of the state, direction, level, and number respectively to obtain encoding units of the state, direction, level, and number; 4. Structural unit coding The state, direction, level and number coding units are combined in the order of state coding unit, direction coding unit, level coding unit and number coding unit to form the structural unit coding of the node.

2. The polymorphic structure encoding method driven by quadruple logic according to claim 1, characterized in that: There are four states: state 1, indicating that the x and y coordinate values ​​of the node are even, even; state 2, indicating that the x and y coordinate values ​​of the node are even, odd; state 3, indicating that the x and y coordinate values ​​of the node are odd, even; state S4, indicating that the x and y coordinate values ​​of the node are odd, odd.

3. The polymorphic structure encoding method driven by quadruple logic according to claim 2, characterized in that: The state coding unit is represented by four digits 1 and 0, the state 1 code is 0001, the state 2 code is 0010, the state 3 code is 0100, and the state 4 code is 1000.

4. The polymorphic structure encoding method driven by quadruple logic according to claim 1, characterized in that: The direction coding unit is represented by four-digit numbers 1 and 0, the right direction code is 0001, the upward direction code is 0010, the downward direction code is 0100, the left direction code is 1000, the inward direction code is 0000, and the outward direction code is 1111.

5. The quaternary logic driven polymorphic structure encoding method according to claim 1, characterized in that: The hierarchical coding unit is represented by four-digit numbers 1 and 0, with the first to fourth layers being represented by four-digit numbers, and the fifth layer and above being represented by at least eight-bit sparse coding.

6. The polymorphic structure encoding method driven by quadruple logic according to claim 1, characterized in that: The numbering and coding unit is selected by one of the following methods: (1) Coding structure: the number field adopts a 4-bit, 8-bit, 12-bit, or 16-bit sparse activation structure, and the bit width is expanded according to the size of the data set and the required coding information dimension; (2) Sparse activation, activating 1 to 4 bits with semantic meaning in the encoding, and keeping the rest at 0; (3) Bit segment grouping, where the numbered bit pattern is divided into at least two functional segments; (4) Symmetrical coding, using mirror image comparison, positive and negative states; (5) Semantic parsing: Each set of activation bits can be reverse parsed through semantic mapping.

7. The quaternary logic driven polymorphic structure encoding method according to claim 1, characterized in that: The number identifier also expresses the multiple semantic mappings of the node in the overall structure of the data set.

8. The quaternary logic driven polymorphic structure encoding method according to claim 7, characterized in that: The multiple semantic mapping is provided with a semantic parsing control table, and the semantic parsing control table is set and determined during encoding.

9. The polymorphic structure encoding method driven by quadruple logic according to claim 8, characterized in that: The multiple semantic mapping switching is performed by selecting one of the following switching methods: (1) Encoding built-in control bit switching Embed a 4-bit structure control code in the high bit of the number field or the reserved field as a switching identifier; (2) Context Binding Bind the semantic control table according to the scenario when calling the external system; (3) External protocol or structure identification trigger Append control instruction flags to structure data flow or task descriptions.

10. An application of the polymorphic structure encoding method driven by quadruple logic as claimed in claims 1 to 9, characterized in that: The quaternary logic-driven polymorphic structure coding method is applied to structure ID card coding, graph neural network GNN structure path input coding, storage and computing integrated chip path coding instructions, data authority and / or experience sharing structure identification or chip circuit coding.

Images