Intellectual property monitoring and early warning method and system based on knowledge graph

Abstract

The invention relates to the technical field of information technology, and discloses an intellectual property monitoring and early warning method and system based on a knowledge graph, and the method comprises the steps: obtaining a technical scheme description inputted by a user, and constructing a target technical path; determining a preset patent set based on the technical function nodes in the target technical path; analyzing the claim text of the preset patent set, extracting technical function elements, technical means elements and technical constraint elements, and constructing a technical path knowledge graph; establishing structural mapping for the target technical path and an alignable claim technical path in the map, keeping the directed connection relation unchanged and keeping the positions of constraint nodes consistent; comparing constraints point by point under structure mapping, determining a constraint partial sequence and judging a dominating relation based on a parameter interval, an execution sequence and exclusive limitation; and when the control is established, judging a path occupation state, identifying an unavoidable structure node, and outputting a structure conflict reasoning result. According to the invention, misjudgment caused by only depending on text matching can be reduced.

Description

Technical Field

[0001] This invention relates to the field of information technology, and more specifically, to a method and system for monitoring and warning intellectual property rights based on knowledge graphs. Background Technology

[0002] As technological innovation activities continue to deepen, various innovation entities are increasingly in need of intellectual property risk assessment during the technology research and development process.

[0003] Current intellectual property monitoring practices mainly rely on patent search systems and text similarity analysis tools to screen potentially related patents through keyword matching or semantic vector calculation. Such methods reveal structural flaws during implementation: First, the comparison between technical solutions and patent claims is limited to surface-level textual features, failing to establish a dynamic correlation model between technical functional nodes, technical means nodes, and technical constraint nodes, resulting in an inability to accurately identify the topological dependencies of each element in the technical path; Second, the inherent multi-layered limiting structure of patent claims is simplified into a discrete set of technical features, and key structural information such as the execution order constraints, parameter range limitations, and exclusivity conditions of technical means are fragmented during analysis, causing a loss of the overall integrity of the technical path; Third, when faced with complex technical solutions with branching and converging structures or necessary nodes, existing technologies lack the ability to deeply analyze path structural conflicts. When the target technical solution has differences in local technical features but the overall path still falls within the scope of the claim structure constraints, it cannot effectively identify unavoidable technical nodes; Fourth, the patent data source screening mechanism relies excessively on a preset classification number system, failing to dynamically construct the analysis scope based on the functional characteristics of the technical solution, resulting in the screening results including a large number of irrelevant patents or omitting key constraint patents.

[0004] These deficiencies render the existing monitoring system fundamentally inadequate in its analytical capabilities at the technical path structure level, making it difficult to accurately determine whether a technical solution falls within the scope of the claims of prior patents, ultimately resulting in a lack of legal reliability support for risk warning results. Summary of the Invention

[0005] In view of this, the present invention proposes an intellectual property monitoring and early warning method and system based on knowledge graph, which aims to solve the problem that relying on text similarity calculation or keyword retrieval for patent comparison makes it difficult to accurately depict the correspondence between the technical solution and the patent claims at the structural level.

[0006] In one aspect, this invention proposes a knowledge graph-based method for monitoring and early warning of intellectual property rights, comprising the following steps: Obtain the user's input description of the technical solution and construct the target technology path; Determine the preset patent set based on the technical functional nodes in the target technology path; Obtain the claim text of a pre-defined patent set, parse it to obtain technical function elements, technical means elements and technical constraint elements, and construct a technical path knowledge graph that includes technical function nodes, technical means nodes, technical constraint nodes, technical function completion nodes and directed connection relationships; For the target technology path and the claim technology paths in the technology path knowledge graph that can be aligned at the level of technology function nodes, a structural mapping is established; the structural mapping maintains the directed connection relationship unchanged and keeps the position of the constraint nodes in the claim technology paths consistent; Under structural mapping, constraints are compared point by point. The constraint partial order is determined according to the parameter interval inclusion relationship, the strong and weak relationship of the execution order and the consistency of the exclusive limitation. When the technical path of the claim satisfies the constraint partial order at all mapping points, it is determined that the technical path of the claim dominates the target technical path. When dominance is established, it is determined whether all technical means nodes of the target technical path fall within the scope of the constraints of the technical path of the claims, and the path occupancy status is marked when it is established; In the technical path of the claims, identify all feasible paths that simultaneously satisfy branch convergence and reach the node where the technical function is completed, and are subject to exclusive constraints or strong dependency constraints, and mark unavoidable structural nodes. The output structure conflict reasoning results are based on path occupancy status and unavoidable structural nodes.

[0007] Furthermore, the pre-defined patent set based on the technical functional nodes in the target technology path includes: Map the technical function nodes in the target technology path to the preset set of technical topic nodes and category nodes in the technology path knowledge graph; Based on the set of technical topic nodes and the set of classification nodes, patents whose publication date is earlier than the generation time of the target technical path and whose technical topic nodes or classification nodes satisfy the mapping relationship are selected from the patent data source to form a preset patent set. After the pre-set patent set is formed, a patent identifier and a claim identifier are established for each patent, and the claim identifier is used as an index key for the technical path of the claim.

[0008] Furthermore, the analysis of the claim text of the pre-defined patent set includes: Sentence segmentation is performed on the claim text and limiting phrases are identified to extract functional description fragments representing the technical purpose as technical functional elements; Identify combinations of action phrases and noun phrases that represent structures, steps, or modules as technical means elements; Identify parameter ranges, sequence words, dependency trigger words, and exclusive qualifiers used to limit the applicable conditions of technical means as technical constraint elements; The technical functional elements, technical means elements, and technical constraint elements are written into the technical path knowledge graph in a three-element structure of functional nodes, means nodes, and constraint nodes, and a constraint type field and a scope field are written for each constraint node.

[0009] Furthermore, establishing a structural mapping that maintains consistency between directed connectivity and the location of constraint nodes includes: Establish a functional anchor mapping between the technical functional nodes of the target technical path and the technical functional nodes of the claim technical path; Under the functional anchor point mapping constraint, the technical means nodes of the target technical path are mapped to the technical means nodes in the claim technical path that have the same pre-order connection structure and subsequent order connection structure according to the directed edge topology order. Map the technical constraint nodes located between adjacent technical means nodes in the target technical path to the technical constraint nodes located between corresponding technical means nodes in the claim technical path; When there are more than two candidate mapping paths, the structure mapping that maximizes the number of constraint node mappings and has zero violations of topological order is selected as the final structure mapping.

[0010] Furthermore, determining the constraint partial order based on the parameter interval inclusion relationship includes: The parameter names of the corresponding parameter range constraint nodes in the target technical path and the claim technical path are normalized and merged with those of the same name; Read the lower and upper limits of the parameter range constraint nodes corresponding to the technical path of the claim, and read the lower and upper limits of the parameter range constraint nodes corresponding to the target technical path; When the lower limit of the parameter interval constraint node corresponding to the technical path of the claim is greater than or equal to the lower limit of the parameter interval constraint node corresponding to the technical path of the target, and the upper limit of the parameter interval constraint node corresponding to the technical path of the claim is less than or equal to the upper limit of the parameter interval constraint node corresponding to the technical path of the target, it is determined that the parameter interval constraint corresponding to the technical path of the claim is not weaker than the parameter interval constraint corresponding to the technical path of the target. When any parameter name is missing an upper or lower limit value in the target technical path, the missing boundary is recorded as an open boundary, and in the constraint partial order determination, only the boundary corresponding to the claim technical path is allowed to be the same open boundary.

[0011] Furthermore, determining the partial order of constraints according to the strength of the execution order includes: Sequence constraint nodes are divided into fixed sequence constraints, partial sequence constraints, and commutative sequence constraints. When the sequential constraint node corresponding to the technical path of the claim is a fixed sequential constraint and the sequential constraint node corresponding to the technical path of the target is a partial sequential constraint or an interchangeable sequential constraint, it is determined that the sequential constraint corresponding to the technical path of the claim is not weaker than the sequential constraint corresponding to the technical path of the target. When the sequential constraint node corresponding to the technical path of the claim is a partial sequential constraint, the constraint of the preceding technical means node in the target technical path is checked based on the directed edge relationship after structural mapping. If the check is successful, the constraint partial order is determined to be satisfied. When a condition-triggered edge appears in the structure mapping, the path segment corresponding to the condition-triggered edge is treated as an independent sequence field, and the order strength relationship is determined separately.

[0012] Furthermore, determining the constraint partial order according to the consistency of exclusive constraints includes: The exclusive constraint node in the technical path of the claim is parsed as a unique mode limitation or a restricted set limitation, and the restricted set limitation is written into the set of allowed implementation modes; Under the structural mapping, read the implementation method identifier of the technical means node corresponding to the target technical path; When the exclusive constraint node is limited by a unique method, the verification target technical path's corresponding implementation method identifier is consistent with the unique method identifier; When the exclusive constraint node is limited by a restricted set, the verification target technical path corresponds to the implementation method identifier that belongs to the set of allowed implementation methods; If the verification fails, it is determined that the partial order constraint of the mapping point is not satisfied and the dominance determination is terminated.

[0013] Furthermore, determining whether the technical means nodes of the target technical path fall within the scope of the constraints of the technical path in the claims includes: Record the starting point technical means node identifier and the ending point technical means node identifier of the constraint's effective range for each technical constraint node; Based on the directed connection relationship, determine the set of reachable path segments from the starting technical means node identifier to the ending technical means node identifier, and define the set of reachable path segments as the scope of constraint; Under structural mapping, each technical means node of the target technical path is mapped to the corresponding technical means node of the claim technical path, and the corresponding technical means node is verified to be within the scope of the constraint. When all technical means nodes of the target technical path meet the verification requirements, the path occupancy status is determined to be valid.

[0014] Furthermore, the reasoning results that identify unavoidable structural nodes and output structures include: A technical means node in the technical path of the claim with a number of preceding directed edges not less than 2 is determined as a branch-merging node; A set of feasible directed paths is constructed with the technical function nodes of the technical path of the claim as the starting point and the technical function completion nodes as the ending point. When any feasible directed path contains the same branch merging node, the branch merging node is determined to be a necessary node. When there are exclusive constraint nodes or dependent constraint nodes among the constraint nodes associated with the necessary node, the necessary node is marked as an unavoidable structural node. The structural conflict reasoning results include at least the dominance establishment marker, the path occupancy status marker, the set of unavoidable structural node identifiers, the mapping path segment identifiers corresponding to the unavoidable structural nodes, and the constraint node identifiers that do not satisfy the constraint partial order.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention no longer relies solely on text similarity or keyword matching. Instead, it models the patent claims and the target technical solution into a unified technical path structure that includes technical function nodes, technical means nodes, technical constraint nodes, and technical function completion nodes. By analyzing the structure mapping and path reasoning, it determines whether the technical solution falls within the scope of the technical path defined by the claims, thereby effectively avoiding misjudgments caused by differences in text expression.

[0016] Based on structural mapping, this invention introduces a constraint partial order determination method that considers parameter interval inclusion relationships, execution order strength relationships, and exclusive constraint consistency. It compares the constraint relationship between the technical solution and the claims point by point, so that the warning results can truly reflect the actual degree of constraint of the claims on the technical solution and avoid the conclusion bias caused by judging only local features.

[0017] By analyzing the branching and converging relationships in the technical path of the claims and the necessary paths to reach the nodes where the technical functions are completed, this invention can identify unavoidable structural nodes that are limited by exclusive constraints or strong dependency constraints. Even if the target technical solution makes adjustments to individual technical means, it can still accurately identify whether there are structural risks in its overall technical path, thereby improving the reliability and interpretability of the early warning results.

[0018] This invention uses the technical functional nodes in the target technical path as functional anchors and dynamically determines the preset patent set through the mapping relationship between technical topics and classifications. This avoids using fixed classification numbers or manual experience rules for patent screening, thereby effectively narrowing the scope of analysis while ensuring the integrity of the analysis and reducing the interference of irrelevant patents on the early warning results.

[0019] By mapping patent claims and technical solutions to a unified technical path knowledge graph and establishing indexes for the technical paths of the claims, this invention provides a clear structural basis and reasoning path for the early warning analysis process, facilitating subsequent tracing, verification, and extended application of early warning results, thereby improving the overall stability and scalability of the system.

[0020] On the other hand, the present invention also provides an intellectual property monitoring and early warning system based on knowledge graphs, comprising: The target path construction module is configured to obtain the technical solution description input by the user, and to analyze the technical function, technical means and technical constraints of the technical solution description to construct the target technical path; The patent set determination module is configured to determine a preset patent set that has a mapping relationship with the technical function nodes in the target technical path from the patent data source. The knowledge graph construction module is configured to obtain the claim text of a preset set of patents, parse out the technical function elements, technical means elements and technical constraint elements, and construct a technical path knowledge graph containing technical function nodes, technical means nodes, technical constraint nodes and their directed connection relationships. The structural mapping module is configured to use technical function nodes as anchors to establish a structural mapping that maintains directed connection relationships and consistent with the position of constraint nodes for the target technical path and the claim technical paths that can be aligned at the technical function node level in the technical path knowledge graph. The dominance relationship determination module is configured to compare constraints point by point under the structural mapping, determine the constraint partial order according to the parameter interval inclusion relationship, the strong and weak relationship of the execution order and the consistency of the exclusive limitation, and determine that the claim technical path dominates the target technical path when the claim technical path satisfies the constraint partial order at all mapping points. The path occupancy determination module is configured to determine whether all technical means nodes of the target technical path fall within the scope of the constraints of the technical path of the claims when the dominance relationship is established, and to mark the path occupancy status when the determination is established. The unavoidable structure identification module is configured to identify, in the technical path of the claim, all feasible paths that simultaneously satisfy branch convergence and reach the node of technical function completion, and are subject to exclusive constraints or strong dependency constraints, and mark the unavoidable structure nodes. The output module is configured to output structural conflict inference results based on path occupancy status and unavoidable structural nodes.

[0021] It is understandable that the aforementioned knowledge graph-based intellectual property monitoring and early warning method and its system have the same beneficial effects, and will not be elaborated further here. Attached Figure Description

[0022] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A flowchart illustrating a knowledge graph-based intellectual property monitoring and early warning method provided in an embodiment of the present invention; Figure 2 This is a functional block diagram of a knowledge graph-based intellectual property monitoring and early warning system provided in an embodiment of the present invention. Detailed Implementation

[0023] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0024] In the field of intellectual property monitoring and early warning technology, existing methods mainly rely on text similarity calculations or keyword searches for patent comparison, which makes it difficult to accurately depict the correspondence between technical solutions and patent claims at the structural level. Specifically, the dependencies, limitations, and execution order relationships among technical functions, technical means, and technical constraints are not included in the analysis framework, resulting in a lack of analytical ability regarding the overall structure of the technical path. Consequently, the process of determining whether a technical solution falls within the claim's defined structure is based solely on matching discrete technical features, failing to reflect the topological connections between nodes in the technical path. This leads to instability and reduced reliability of risk assessment results, directly affecting the accuracy of intellectual property risk warnings.

[0025] For example, in the development of semiconductor chip manufacturing processes, companies propose a novel photolithography alignment technology, involving technical functional nodes such as exposure energy control, mask positioning, and defect compensation. Existing monitoring systems, after filtering relevant patents based on preset classification numbers, only perform fragmented comparisons of technical features, ignoring the strong dependency between energy parameter thresholds and positioning steps in the technical path. When the target technical solution differs in local parameter settings, the system incorrectly determines that there is no conflict risk, while in reality, the overall technical path still falls within the structural constraint space defined by prior patent claims, resulting in the failure to effectively identify structural infringement risks at key process nodes.

[0026] If the aforementioned issues are not addressed, the intellectual property monitoring and early warning system will be unable to identify unavoidable key technical nodes in the technical solutions, leading to potential patent infringement uncertainties for enterprises during the technology implementation phase. This could cause R&D resources to be invested in technical paths with structural risks, or trigger legal disputes after product commercialization, thereby hindering the technological innovation process and affecting market competitiveness.

[0027] For this, please refer to Figure 1 As shown, this application proposes a knowledge graph-based intellectual property monitoring and early warning method, including: S1: Obtain the user's input description of the technical solution and construct the target technical path; S2: Determine the preset patent set based on the technical functional nodes in the target technology path; S3: Obtain the claim text of the preset patent set, parse it to obtain the technical function elements, technical means elements and technical constraint elements, and construct a technical path knowledge graph that includes technical function nodes, technical means nodes, technical constraint nodes, technical function completion nodes and directed connection relationships; S4: Establish a structural mapping between the target technical path and the claim technical paths in the knowledge graph that can be aligned at the level of technical function nodes; the structural mapping maintains the directed connection relationship unchanged and keeps the position of the constraint nodes in the claim technical paths consistent; S5: Under structural mapping, compare constraints point by point, determine the constraint partial order according to the parameter interval inclusion relationship, execution order strength relationship and exclusive limitation consistency, and determine that the claim technical path dominates the target technical path when the claim technical path satisfies the constraint partial order at all mapping points. S6: When dominance is established, determine whether all technical means nodes of the target technical path fall within the scope of the constraints of the technical path of the claim, and mark the path occupancy status when it is established; S7: Identify, in the technical path of the claim, all feasible paths that simultaneously satisfy branch convergence and reach the node where the technical function is completed, and which are subject to exclusive constraints or strong dependency constraints, and mark the unavoidable structural nodes. S8: Output structural conflict inference results based on path occupancy status and unavoidable structural nodes.

[0028] For ease of understanding, the following explains some key terms in this embodiment: Technical solution description: This refers to the technical content provided by the user in natural language or structured text form that requires intellectual property risk assessment. This description typically includes the technical objective, the technical means employed to achieve that objective, and any potential limitations.

[0029] The target technology path is a graph representation formed after the structured analysis of the technical solution description. This path consists of a series of technical function nodes, technical means nodes, and technical constraint nodes, as well as the directed connections between them, used to clearly express the logical flow and structural composition of the technical solution.

[0030] Technical function node: This refers to the part of a technical solution that achieves a specific technical purpose or function. For example, in a "data processing" solution, "data input," "data conversion," and "data output" can all be considered technical function nodes.

[0031] Technical means nodes: These represent the specific structures, components, steps, or modules used in a technical solution to achieve the technical function. For example, under the "data input" function, there may be technical means nodes such as "collecting data through sensors" or "receiving data through a user interface".

[0032] Technical constraint nodes: These represent the limiting conditions imposed on technical means or functions, such as parameter ranges, execution order, dependencies, or exclusivity requirements. These constraint nodes are used to precisely define the applicable conditions and implementation methods of technical solutions.

[0033] Technical Function Completion Node: This node indicates that all technical functions in the technical solution have been implemented, and the entire technical process has reached its expected endpoint.

[0034] The technology path knowledge graph is a large-scale knowledge base that stores a wealth of structured information about patent claims. This graph represents technical function nodes, technical means nodes, technical constraint nodes, and the directed connections between them in graph form, aiming to provide a set of patent technology structures for comparison and analysis.

[0035] The claim technical path is a structured representation of a specific patent, extracted from a technical path knowledge graph. It also consists of technical function nodes, technical means nodes, technical constraint nodes, and directed connections, used for structured comparison with the target technical path.

[0036] Structural mapping is a correspondence established between the target technical path and the technical path of the claims. This mapping aims to identify corresponding points in the two paths at the levels of technical function, technical means, and technical constraints, and to maintain the consistency of the directed connection relationship and the position of the constraint nodes between them.

[0037] Constraint partial order: Under structural mapping, this refers to the priority or subordination relationship determined by comparing the corresponding constraint nodes in the target technical path and the claim technical path. This partial order is used to determine whether the constraints of the claim technical path are stronger than, or at least not weaker than, the constraints of the target technical path.

[0038] Domination: This refers to the situation where the technical path of the claim satisfies the constraint partial order at all mapping points, indicating that the technical path of the claim has a dominant relationship with the target technical path. This usually means that the target technical path falls within the protection scope of the technical path of the claim.

[0039] Path occupancy status: This refers to determining whether all technical means nodes of the target technical path are covered by the scope of the claims' technical path, assuming the dominance relationship is established. If all nodes are covered, the path occupancy status is established.

[0040] Unavoidable structural nodes: These refer to technical means nodes in the technical path of the claims that simultaneously satisfy the requirements of branch convergence, all feasible paths leading to the node of technical function completion, and are subject to exclusive or strong dependency constraints. These nodes typically represent the core technical features of the claims and are difficult to circumvent through simple modifications.

[0041] Structural conflict reasoning results: These are risk assessment conclusions output based on path occupancy status and unavoidable structural nodes analysis. These results indicate whether a structural conflict exists between the target technical solution and prior patent claims, and the specific manifestation of the conflict.

[0042] This embodiment provides a knowledge graph-based method for intellectual property monitoring and early warning. This method achieves accurate assessment of intellectual property risks by performing structured analysis of technical solutions and comparing them with the structured representation of patent claims.

[0043] First, the method includes obtaining a user-input description of the technical solution and constructing a target technical path. Users can input the technical solution description in various ways, such as directly inputting a natural language description through a text editor, or filling in information about technical functions, technical means, and technical constraints through a structured form. After obtaining the technical solution description, it needs to be parsed to identify the technical functions, technical means, and technical constraints, and to establish the logical relationships between them, ultimately forming a directed graph structure of the target technical path. For example, a technical solution describing "collecting temperature data through a sensor, then analyzing the data through a processor, and finally displaying the results on a display" can be parsed into a path structure of "collecting temperature data" (technical function node) -> "sensor" (technical means node) -> "data analysis" (technical function node) -> "processor" (technical means node) -> "displaying results" (technical function node) -> "display" (technical means node).

[0044] Furthermore, a pre-defined patent set is determined based on the technical functional nodes within the target technical path. After constructing the target technical path, the technical functional nodes within it are identified. These technical functional nodes can be used for preliminary screening of relevant patent documents. For example, a search can be performed in a patent database based on the keywords or semantic information of the technical functional nodes to obtain patents related to these functions. Another approach is to pre-establish a mapping relationship between technical functions and patent classification numbers or technical topics, and then screen a preliminary patent set by matching the technical functional nodes in the target technical path with these pre-defined classifications or topics.

[0045] Subsequently, the method obtains the claim text of a pre-defined patent set, parses it to obtain technical functional elements, technical means elements, and technical constraint elements, and constructs a technical path knowledge graph containing technical functional nodes, technical means nodes, technical constraint nodes, technical function completion nodes, and directed connections. After obtaining the pre-defined patent set, the claim text of each patent needs to be analyzed in depth. This includes word segmentation, syntactic analysis, and semantic understanding of the claim text to accurately identify the technical functions, technical means, and technical constraints contained therein. For example, phrases expressing "for achieving function A" in the claim can be identified as technical functional elements, phrases such as "through module B" or "perform step C" as technical means elements, and phrases such as "under condition D" or "in manner E" as technical constraint elements. These parsed elements will be stored in a structured manner, and directed connections will be established between them, thus forming a large technical path knowledge graph. This knowledge graph not only contains each technical element but also clarifies their logical dependencies and execution order, and marks the technical function completion nodes.

[0046] Based on this, a structural mapping is established between the target technical path and the claim technical paths in the knowledge graph that can be aligned at the technical function node level. This structural mapping maintains the directed connection relationship and ensures that the position of the constraint nodes in the claim technical paths remains consistent. Structural mapping is the core step of the comparison. First, technical function nodes with the same or similar semantics in the target technical path and claim technical paths are identified and used as anchor points for alignment. For example, if there is a "data acquisition" function node in the target technical path and a "data gathering" function node in the claim technical path, a functional anchor point mapping can be established between them. Under the constraint of this functional anchor point, further attempts are made to map the technical means nodes and technical constraint nodes in the target technical path to the corresponding nodes in the claim technical path. This mapping process must strictly follow the directed connection relationship to ensure that the path structure after mapping remains consistent with the original path structure, and that the position of the technical constraint nodes does not change before and after mapping.

[0047] Furthermore, constraints are compared point-by-point under the structural mapping. The constraint partial order is determined according to the inclusion relationship of parameter ranges, the strength of execution order, and the consistency of exclusive constraints. When the claim technical path satisfies the constraint partial order at all mapped points, it is determined that the claim technical path dominates the target technical path. After establishing the structural mapping, a detailed constraint comparison is performed on the corresponding mapped nodes. For example, if the target technical path has a parameter constraint "temperature range of 20-30 degrees Celsius," while the corresponding parameter constraint in the claim technical path is "temperature range of 25-28 degrees Celsius," then the constraint of the claim technical path is considered no weaker than the constraint of the target technical path because its parameter range is included in the parameter range of the target technical path. As another example, if the execution order of two technical means in the target technical path is "A then B or B then A," while the corresponding order in the claim technical path is "must A then B," then the order constraint of the claim technical path is considered stronger than the commutative order constraint of the target technical path. In addition, for exclusive constraints, it is verified whether the implementation method of the target technical path is consistent with the exclusive constraint of the claim technical path. Only when the constraints of the technical path in the claim satisfy the partial order relation at all mapping points is it determined that the technical path in the claim has a dominant relationship with the target technical path.

[0048] When dominance is established, it is determined whether all technical means nodes of the target technical path fall within the binding scope of the claim technical path. If so, the path occupancy status is marked. The establishment of dominance only indicates that the constraints of the claim technical path are no weaker than those of the target technical path. To more accurately assess risk, it is necessary to further determine whether all technical means nodes in the target technical path are covered by the constraints of the claim technical path. For example, a constraint in the claim technical path may only apply to a specific set of technical means nodes within its path. The scope of these constraints will be identified, and it will be checked whether all corresponding technical means nodes in the target technical path fall within these scopes. If all technical means nodes are covered, the path occupancy status is marked, indicating that the target technical solution is also fully defined by the claim at the technical means level.

[0049] Furthermore, in the technical path of the claims, nodes that simultaneously satisfy the conditions of branch convergence, are necessary for all feasible paths to reach the technical function completion node, and are subject to exclusive or strong dependency constraints are identified, and these nodes are marked as unavoidable structural nodes. To identify core risk points, the topological structure of the technical path of the claims is analyzed. For example, nodes with multiple preceding directed edges are identified as "branch convergence" nodes. Then, from the starting functional node of the technical path of the claims to the technical function completion node, all possible paths are traversed to find those "necessary nodes" that all feasible paths must pass through. If these necessary nodes are simultaneously subject to exclusive or strong dependency constraints, they are marked as "unavoidable structural nodes." These nodes represent core technical features in the claims that are difficult to circumvent, posing a substantial risk to the target technical solution.

[0050] Finally, the structural conflict inference result is output based on the path occupancy status and unavoidable structural nodes. The final structural conflict inference result is generated by combining the path occupancy status and the identification results of unavoidable structural nodes. This result may include the dominance assertion marker, the path occupancy status marker, the set of identifiers of the identified unavoidable structural nodes, and the identifiers of the corresponding mapped path segments for these unavoidable structural nodes. Furthermore, if there are constraint nodes that do not satisfy the constraint partial order, their identifiers will also be included in the inference result. This information provides users with a detailed intellectual property risk assessment report, indicating potential conflict points and the degree of risk.

[0051] The following specific example will provide a more in-depth explanation of the above technical solution: Suppose user A is developing a new type of "intelligent environmental monitoring system" and has entered its technical solution description: "This system collects air quality data through sensors, then transmits the data to a cloud server for processing. The server determines whether the air quality meets the standards based on preset thresholds. If it does not meet the standards, it sends a warning notification to the user through a mobile application." First, obtain the description of the technical solution input by user A, and construct the target technical path. This description is parsed into the following structure: Technical functional nodes: data acquisition, data transmission, data processing, and early warning notification.

[0052] Technical nodes: sensors (corresponding to data acquisition), cloud servers (corresponding to data transmission and data processing), and mobile applications (corresponding to early warning notifications).

[0053] Technical constraint node: preset threshold (limits data processing), if the threshold is not met, a warning notification will be sent (limits the triggering conditions for warning notification).

[0054] This forms a directed connection target technology path.

[0055] Secondly, based on the technical functional nodes (data acquisition, data transmission, data processing, and early warning notification) in the target technology path, a pre-defined patent set is determined. Relevant patents are then searched in the patent data source according to these functional nodes. For example, patents related to topics such as "environmental monitoring," "data transmission," and "intelligent early warning" may be retrieved, forming a preliminary pre-defined patent set.

[0056] Next, the claim texts of the pre-defined patent set are obtained and parsed to construct a technology path knowledge graph. For example, the claim text of a prior patent P1 is obtained from the pre-defined patent set. This claim describes an "intelligent environmental data processing method," including "acquiring environmental parameters through environmental sensors, uploading the acquired environmental parameters to a remote server, analyzing the environmental parameters on the remote server, and pushing abnormal information to the user via a mobile terminal when the analysis results exceed the safe range." This claim is parsed into a claim technology path, which includes corresponding technical function nodes, technical means nodes, technical constraint nodes, and directed connection relationships, and is incorporated into the technology path knowledge graph.

[0057] Based on this, a structural mapping is established between User A's target technical path and the technical path of Patent P1's claims. First, functional anchor points are identified, such as "data acquisition" and "acquiring environmental parameters," "data transmission" and "uploading to a remote server," "data processing" and "analyzing environmental parameters," and "early warning notification" and "pushing abnormal information." Under the constraints of these functional anchor points, "sensor" in the target technical path is mapped to "environmental sensor" in the claim technical path, "cloud server" is mapped to "remote server," and "mobile application" is mapped to "mobile terminal." Simultaneously, the "preset threshold" constraint in the target technical path is mapped to the "out of safe range" constraint in the claim technical path, and the "send if not met" constraint is mapped to the trigger condition constraint "pushing abnormal information to the user via mobile terminal when the analysis result exceeds the safe range." This mapping maintains consistency in the directed connection relationships and the positions of the constraint nodes.

[0058] Subsequently, constraints are compared point-by-point under structural mapping to determine the partial order of constraints. Specifically: Regarding parameter range constraints: Suppose that the "preset threshold" in the target technical path is limited to "PM2.5 concentration higher than 50 ug / m³", while the "exceeding the safe range" in the claim technical path is limited to "PM2.5 concentration higher than 60 ug / m³". In this case, the constraint of the claim technical path (higher than 60) is determined to be no weaker than the constraint of the target technical path (higher than 50), because the condition of the target technical path (higher than 50) includes the condition of the claim technical path (higher than 60).

[0059] Regarding execution order constraints: Assume that the order of "data acquisition" and "data transmission" in the target technical path is interchangeable, while the claim technical path explicitly specifies "acquiring environmental parameters first, then uploading them to the remote server." In this case, the fixed order constraint of the claim technical path is determined to be stronger than the interchangeable order constraint of the target technical path.

[0060] Regarding the consistency of exclusive limitation: Suppose that the implementation of the "sensor" in the target technical path is a "laser scattering sensor", while the corresponding "environmental sensor" in the claim technical path does not have an exclusive limitation. In this case, the consistency of exclusive limitation is satisfied.

[0061] When the constraint partial order of all mapping points is satisfied, it is determined that the technical path of the claim of patent P1 dominates the target technical path of user A.

[0062] When dominance is established, it is determined whether all technical means nodes of the target technical path fall within the scope of the constraints of the technical path in the claims. For example, in the claim of patent P1, the constraint "when the analysis result exceeds the safe range" covers the two technical means of "remote server analyzing environmental parameters" and "pushing abnormal information to users through mobile terminals". It is verified whether the "cloud server" and "mobile application" corresponding to user A's target technical path both fall within the scope of this constraint. If both do, the path occupancy status is marked as established.

[0063] Furthermore, unavoidable structural nodes are identified in the technical path of the claims of patent P1. For example, if the technical means node of "remote server analyzing environmental parameters" is a node that all feasible paths from "acquiring environmental parameters" to "pushing abnormal information" must pass through in the technical path of the claims of patent P1, and this node is subject to the strong dependency constraint of "when the analysis result exceeds the safe range", then the technical means node of "remote server analyzing environmental parameters" will be marked as an unavoidable structural node.

[0064] Finally, based on the path occupancy status and unavoidable structural nodes, the structural conflict inference result is output. This result clearly indicates that User A's "Intelligent Environmental Monitoring System" technical solution is dominated by the technical path of the claims in Patent P1, and the path occupancy status is established. Simultaneously, "remote server analyzing environmental parameters" is identified as an unavoidable structural node, meaning that User A's solution has a high degree of overlap with Patent P1 in its core processing steps, and this overlap is difficult to avoid.

[0065] Based on the above examples, the method provided in this embodiment demonstrates a significant technical contribution in solving existing technical problems. Existing technologies primarily rely on keyword retrieval or text similarity analysis to assess intellectual property risks; for example, comparing the textual similarity between user A's technical solution description and patent P1 to determine relevance. This method struggles to accurately capture the deep structural relationship between the technical solution and the claims.

[0066] In contrast, this embodiment constructs a knowledge graph of the target technology path and the technology path, transforming both the technical solution and the claims into structured graph representations. This structured representation can accurately depict the directed connections between technical functions, technical means, and technical constraints, overcoming the problem of existing technologies that decompose claims into discrete technical features while ignoring their inherent logical structure. For example, in the above example, this method can clearly identify the function-means relationship between "data acquisition" and "sensor," as well as the limiting effect of "preset threshold" on "data processing," rather than simply identifying the words "sensor" and "threshold."

[0067] Furthermore, this method achieves precise determination of the dominance relationship between the technical solution and the claims by establishing structural mapping and point-by-point comparison constraints. Existing methods often lead to unstable or overly coarse judgments when determining whether a technical solution falls within the scope of the claims due to a lack of detailed comparison of constraints. This method, however, uses multiple dimensions such as parameter range inclusion relationships, execution order strength relationships, and consistency of exclusive constraints to determine the constraint order, ensuring the rigor and accuracy of the comparison. For example, in the example, this method can identify the dominance relationship between the constraint "PM2.5 concentration higher than 60ug / m³" in patent P1 and "PM2.5 concentration higher than 50ug / m³" in user A's solution. This detailed constraint comparison is difficult to achieve with traditional text similarity methods.

[0068] Furthermore, this method introduces the identification of path occupancy status and unavoidable structural nodes, providing a deeper level of insight for intellectual property risk assessment. Existing technologies lack the ability to analyze the overall structure of technical paths, making it difficult to identify whether there are unavoidable key technical nodes in the technical solution. This method, by identifying branch convergence, mandatory nodes, and technical means nodes limited by exclusive or strong dependency constraints, can accurately mark the core risk points in the claims. In the example above, identifying "remote server analyzing environmental parameters" as an unavoidable structural node can intuitively prompt user A that the core processing logic of their solution has an unavoidable structural conflict with prior patents, thereby effectively improving the reliability of the warning results.

[0069] In summary, the method provided in this embodiment, through structured representation, refined constraint comparison, and identification of core risk points, can accurately depict the relationship between the technical solution and prior patent claims at the technical path structure level, achieve accurate analysis of intellectual property risks, provide a reliable data foundation for subsequent risk assessment, and significantly improve the accuracy and depth of intellectual property monitoring and early warning.

[0070] In knowledge graph-based intellectual property monitoring and early warning methods, a pre-defined set of patents needs to be established as the basis for analysis. However, if the method for determining the pre-defined patent set is not precise or efficient enough, the subsequent knowledge graph construction and control determination process may face problems such as excessive data volume, insufficient relevance, or omission of key patents, thereby affecting the accuracy and efficiency of the early warning.

[0071] In response, this application further proposes a step for determining a preset patent set based on the technical functional nodes in the target technical path, specifically including: mapping the technical functional nodes in the target technical path to a preset set of technical topic nodes and a set of classification nodes in the technical path knowledge graph; based on the set of technical topic nodes and the set of classification nodes, selecting patents from the patent data source whose publication date is earlier than the time of generation of the target technical path and whose technical topic nodes or classification nodes satisfy the mapping relationship, to form a preset patent set; after the preset patent set is formed, establishing a patent identifier and a claim identifier for each patent, and using the claim identifier as the index key of the claim technical path.

[0072] This involves mapping the technical function nodes in the target technical path to a pre-defined set of technical topic nodes and category nodes in the technical path knowledge graph. The aim is to establish a connection between the user's technical solution and the existing technical classification system. This can be achieved in several ways. For example, natural language processing techniques can be used to extract keywords or phrases from the descriptions of the technical function nodes in the target technical path, and then perform semantic matching or ontology-based mapping with the pre-defined technical topic nodes and category nodes in the technical path knowledge graph. Another approach is to train a machine learning model that can automatically categorize technical function nodes into the most relevant technical topic and category nodes based on their textual content.

[0073] Based on the aforementioned set of technical topic nodes and classification nodes, selecting patents from the patent data source whose publication date is earlier than the generation time of the target technical path and whose technical topic nodes or classification nodes satisfy the mapping relationship, forming a preset patent set, is a crucial step in ensuring the relevance and timeliness of the selected patents. In practice, a query request can be sent to one or more patent databases (e.g., the State Intellectual Property Office patent database, the European Patent Office database, etc.). The query conditions include the mapped technical topic nodes and classification nodes, and stipulate that the publication date of the patent must be earlier than the generation time of the target technical path. For example, a Boolean query can be constructed to connect all relevant technical topics and classification nodes using an "OR" logic and connect them with the time constraint using an "AND" logic to accurately filter out patent documents that meet the conditions.

[0074] After the pre-defined patent set is formed, a patent identifier and a claim identifier are created for each patent. The claim identifier serves as an index key for the claim's technical path, enabling efficient management and access to this patent information later. The patent identifier is typically the patent's publication number or application number, possessing global uniqueness. The claim identifier can be based on the patent identifier with an appended suffix indicating the claim sequence number; for example, "Patent No._C1" represents the first claim of the patent. These identifiers, acting as index keys, allow for rapid location of specific claim texts and their parsing into a technical path structure when constructing a technical path knowledge graph.

[0075] This method precisely maps the technical function nodes in the user-input technical solution description to a pre-defined set of technical topic nodes and classification nodes in a technology path knowledge graph, thus providing a clear and focused search scope for subsequent patent screening. Based on these precisely mapped nodes, patent documents highly relevant to the target technology path and meeting the publication date requirements are efficiently screened from massive patent data sources, avoiding the interference of a large amount of irrelevant data that may result from traditional search methods. Subsequently, a unique identifier is established for each selected patent and its claims, and the claim identifier is used as the index key for the claim technology path. This not only ensures data traceability but, more importantly, provides an efficient data access mechanism for subsequently parsing the claim text and constructing a technology path knowledge graph. This systematic process of determining the pre-defined patent set significantly improves the accuracy and efficiency of the initial stage of intellectual property monitoring and early warning, laying a solid foundation for subsequent structural mapping, dominance determination, and conflict reasoning.

[0076] In some of the embodiments described above in this application, a method for determining a preset patent set based on the technical function nodes in the target technical path is proposed. However, when parsing the claim text of the preset patent set, if a structured method is not used, it is difficult to accurately extract the technical function elements, technical means elements, and technical constraint elements from the complex natural language text, thereby affecting the construction quality of the technical path knowledge graph and the accuracy of subsequent early warning analysis.

[0077] In response, this application further proposes a specific method for parsing the claim text of a pre-defined patent set. This method includes: performing sentence segmentation on the claim text and identifying limiting phrases; extracting functional description fragments representing the technical purpose as technical functional elements; identifying action phrases and noun phrase combinations representing structures, steps, or modules as technical means elements; identifying parameter ranges, sequence words, dependency trigger words, and exclusive limiting words used to limit the applicable conditions of technical means as technical constraint elements; and writing the technical functional elements, technical means elements, and technical constraint elements into a technical path knowledge graph according to a three-element structure of functional nodes, means nodes, and constraint nodes, and writing constraint type fields and scope fields for each constraint node.

[0078] Specifically, sentence segmentation of the claims text aims to break down complex, long sentences into independent, easily analyzable shorter sentences. This can be achieved through rule matching based on punctuation (e.g., periods, semicolons, commas) or through syntactic analysis techniques in Natural Language Processing (NLP), such as using a dependency parser to identify sentence boundaries. Building upon sentence segmentation, defining phrases are identified, which typically contain modifying or limiting information about the technical features. For example, part-of-speech tagging and chunking techniques can be used to extract noun phrases, verb phrases, etc., or deep learning models can be used to identify key phrases in the text.

[0079] Extracting functional descriptive fragments representing the technical purpose as technical functional elements refers to identifying the parts of text that describe the purpose or effect that the invention intends to achieve from the text after sentence segmentation and limiting phrase identification. This typically involves verbs or verb phrases, such as "used to achieve X" or "able to accomplish Y". This process can employ rule-based methods based on keyword matching, such as pre-setting a list of functional verbs, or utilize machine learning classifiers, such as Support Vector Machines (SVM) or Recurrent Neural Networks (RNN), to classify text fragments to identify functional descriptions.

[0080] Identifying combinations of action phrases and noun phrases representing structures, steps, or modules as technical means elements refers to identifying the specific ways in which technical functions are implemented. Technical means are typically manifested as a certain structure, execution steps, or used modules. For example, part-of-speech tagging can be used to identify verbs (actions) and nouns (structures, modules), and then dependency parsing or semantic role labeling can be used to identify the agent-patient relationship between verbs and nouns, thereby combining them into technical means elements. Another approach is to use pre-trained language models (such as BERT, GPT) for entity recognition and relation extraction to identify phrases representing technical means.

[0081] Identifying parameter ranges, sequence words, dependency triggers, and exclusive qualifiers used to limit the applicability of technical means as technical constraint elements refers to identifying the conditions that limit technical means. Parameter ranges refer to the limitation of numerical parameters, such as "temperature between 100℃ and 200℃," which can be identified by matching numbers, units, and comparison operators using regular expressions. Sequence words indicate the order of operations or steps, such as "first" or "then," and can be identified by keyword matching or sequence labeling models. Dependency triggers indicate that the occurrence of an operation or state depends on other conditions, such as "when..." or "if...then," and can be identified by conditional clause pattern matching or semantic dependency analysis. Exclusive qualifiers indicate that a certain method is unique or excludes other methods, such as "limited to" or "uniquely," and can be identified by keyword matching or context analysis. The identification of these qualifiers can combine lexical analysis, syntactic analysis, and semantic analysis techniques; for example, constructing a dictionary containing various qualifiers and combining them with contextual information for matching and judgment.

[0082] Technical functional elements, technical means elements, and technical constraint elements are written into the technology path knowledge graph in a ternary structure of functional nodes, means nodes, and constraint nodes. Each constraint node is assigned a constraint type field and a scope field. This means that the extracted technical functional elements, technical means elements, and technical constraint elements are respectively used as functional nodes, means nodes, and constraint nodes in the knowledge graph. These nodes are connected by directed edges, forming a logical relationship of "function-means-constraint." For example, a functional node can be connected to one or more means nodes, and a means node can be defined by one or more constraint nodes. The constraint type field records the specific type of constraint, such as "parameter range constraint," "sequence constraint," "exclusive constraint," and "dependency constraint," which helps in subsequent classification and comparison during dominance determination. The scope field records the range of technical means nodes affected by the constraint node. For example, a constraint may only affect adjacent means nodes or it may affect a continuous sequence of means nodes. This can be determined by analyzing the position of the constraint in the claim text and the components it modifies. For example, the starting and ending means node identifiers of the constraint's effect are recorded.

[0083] This application's solution systematically analyzes the claim text of a pre-defined patent set, transforming the originally unstructured natural language text into structured technical function nodes, technical means nodes, and technical constraint nodes, and then incorporating them into a technology path knowledge graph in a ternary structure. This structured processing not only clarifies the directed connections between technical elements but also adds constraint type and scope fields to each constraint node, greatly enhancing the expressive power and analyzability of the knowledge graph. In this way, subsequent structural mapping and dominance determination can be based on clear and explicit structured data, overcoming the ambiguity and complexity of directly processing natural language text, and providing a solid data foundation for intellectual property monitoring and early warning.

[0084] The following is a concrete example to illustrate this. Suppose there is a claim text that reads: "A processing method for improving image sharpness, comprising: first, acquiring an original image; then, in an image processing unit, performing Gaussian filtering on the original image, wherein the kernel size of the Gaussian filter is not less than 3x3; next, performing edge enhancement on the filtered image, wherein the edge enhancement uses only the Sobel operator; and finally, outputting the processed image." First, the claim text is segmented into multiple independent short sentences. Next, limiting phrases such as "for improving image sharpness," "the kernel size of the Gaussian filter is not less than 3x3," and "using only the Sobel operator" are identified. Then, the functional description fragment representing the purpose of the technique, "for improving image sharpness," is extracted as a technical functional element. Combinations of action phrases and noun phrases representing structures, steps, or modules, such as "acquiring an original image," "performing Gaussian filtering on the original image," "performing edge enhancement on the filtered image," and "outputting the processed image," are identified as technical means elements. Simultaneously, parameter ranges, sequence words, dependent trigger words, and exclusive qualifiers used to limit the applicability of technical means are identified as technical constraint elements, such as "first," "then," "the kernel size of Gaussian filtering is not less than 3x3," "next," "only the Sobel operator is used," and "finally." Finally, these technical functional elements, technical means elements, and technical constraint elements are written into the technical path knowledge graph according to a ternary structure of functional nodes, means nodes, and constraint nodes. For example, the functional node is "improve image clarity," the means nodes are, in sequence, "acquire the original image," "Gaussian filtering," "edge enhancement," and "output the processed image." The constraint type of the constraint node "first" is a sequence constraint, and its scope is the "acquire the original image" means node; the constraint type of the constraint node "the kernel size of Gaussian filtering is not less than 3x3" is a parameter range constraint, and its scope is the "Gaussian filtering" means node; the constraint type of the constraint node "only the Sobel operator is used" is an exclusive constraint, and its scope is the "edge enhancement" means node.

[0085] Through the above technical solution, this application can transform unstructured claim text into a structured knowledge graph representation, significantly improving the accuracy and completeness of technical element extraction. This refined parsing and structured storage enables subsequent structural mapping and dominance determination to be based on clear and explicit semantic information, thereby effectively avoiding misjudgments caused by text comprehension biases, improving the accuracy and reliability of intellectual property monitoring and early warning, and providing users with more valuable early warning results.

[0086] In some of the above embodiments, a method for constructing a technology path knowledge graph and structurally mapping the target technology path to the claim technology path for dominance determination is proposed. However, in practice, the complexity of technology paths and the various possible connections between nodes may lead to challenges in the structural mapping process, particularly in how to accurately and efficiently establish mapping relationships, and especially in selecting the optimal mapping to ensure the accuracy and reliability of subsequent dominance determination when multiple potential mapping paths exist.

[0087] In response, this application further proposes a method for establishing a structural mapping that maintains the consistency between directed connection relationships and constraint node positions. This method includes: establishing a functional anchor mapping between the technical functional nodes of the target technical path and the technical functional nodes of the claim technical path; under the constraint of the functional anchor mapping, mapping the technical means nodes of the target technical path to the technical means nodes in the claim technical path that have the same pre-order connection structure and subsequent order connection structure according to the topological order of the directed edges; mapping the technical constraint nodes in the target technical path located between adjacent technical means nodes to the technical constraint nodes in the claim technical path located between corresponding technical means nodes; when there are two or more candidate mapping paths, the structural mapping that maximizes the number of constraint node mappings and has zero topological order violations is preferentially selected as the final structural mapping.

[0088] Specifically, establishing functional anchor mapping between the technical functional nodes of the target technical path and the technical functional nodes of the claim technical path aims to provide a stable starting point and reference system for subsequent mapping of more granular technical means nodes and technical constraint nodes, ensuring semantic consistency of the mapping. This mapping can be established by comparing the textual descriptions and semantic similarity of technical functional nodes, or through predefined synonyms, hypernyms, and hyponyms. Alternatively, it can be achieved by manually annotating or using a pre-defined functional classification system from an expert system to match functional nodes with the same classification labels.

[0089] Under the functional anchor point mapping constraint, the technical means nodes of the target technical path are mapped to technical means nodes in the claim technical path that have the same preceding and succeeding connection structures according to the topological order of directed edges. This aims to ensure that the mapping of technical means nodes not only considers their own functions, but more importantly, their contextual relationships in the entire technical path, i.e., their predecessors and successors. This can be achieved by using a graph traversal algorithm, starting from the functional anchor point, gradually matching technical means nodes along the directed edges, and checking whether their direct predecessors and successors have corresponding mapping relationships in another path. Alternatively, it can be achieved by calculating the relative position or level of the technical means node in the path, combined with the type or identifier of its predecessor and successor nodes, to construct a local structural fingerprint for matching.

[0090] Mapping technical constraint nodes located between adjacent technical means nodes in the target technical path to corresponding technical constraint nodes located between corresponding technical means nodes in the claim technical path aims to ensure the contextual correctness of constraints, i.e., that a constraint is specific to a particular technical means or sequence of technical means. After mapping the technical means nodes, the mapped pairs of technical means nodes in the target technical path can be traversed to check if there are technical constraint nodes between them, and matching technical constraint nodes can be found among the corresponding mapped pairs of technical means nodes in the claim technical path. Matching can be based on constraint type, parameter name, or semantic similarity.

[0091] When there are more than two candidate mapping paths, the final mapping is selected by prioritizing the structure that maximizes the number of constraint nodes mapped and has zero topological order violations, aiming to resolve mapping ambiguities. A score can be calculated for each candidate mapping path, with the scoring criteria including the number of constraint nodes mapped (score) and the number of topological order violations (deduction). The path with the highest score is selected. The number of topological order violations can be calculated by comparing whether the relative order of nodes changes before and after the mapping.

[0092] This application's solution first establishes a functional anchor mapping, laying a semantic foundation for structural mapping. Then, under the constraints of the functional anchors, it precisely matches technical means nodes by considering the topological order of directed edges and the preceding and following connection structures, ensuring the logical coherence of the mapping. Subsequently, it maps technical constraint nodes to corresponding technical means nodes, ensuring the contextual correctness of the constraints. Finally, by setting explicit priority rules—maximizing the number of constraint node mappings and minimizing the number of topological order violations—it resolves potential mapping ambiguities in complex technical paths. This hierarchical, progressive mapping method with an optimization selection mechanism ensures the accuracy and completeness of structural mapping, providing a solid foundation for subsequent constraint comparison and dominance determination.

[0093] The following is a concrete example to illustrate this. Assume the target technical path is "Function A -> Means M1 (Constraint C1) -> Means M2 -> Function B", and the claim technical path contains two possible mapping objects: "Function A' -> Means M1' (Constraint C1') -> Means M2' -> Function B'" and "Function A'' -> Means M1'' -> Means M2'' (Constraint C2'') -> Function B''". First, identify whether "Function A" in the target technical path is functionally aligned with "Function A'" and "Function A''" in the claim technical path, and also whether "Function B" is aligned with "Function B'" and "Function B''", thus establishing a functional anchor mapping. Based on this, attempt to map "Means M1" in the target technical path to "Means M1'" or "Means M1''" in the claim technical path. For example, if the preorder connection structure of "means M1" is "function A" and the postorder connection structure is "means M2", then we search for "means M1'" in the claim's technical path that also has "function A'" as the preorder and "means M2'" as the postorder for mapping. Next, we identify the "constraint C1" between "means M1" and "means M2" in the target technical path, and find the corresponding "constraint C1'" between "means M1'" and "means M2'" in the claim's technical path for mapping. Assume that during the mapping process, two candidate mapping paths are found: path 1 maps 3 technical means nodes and 2 technical constraint nodes, and the topological order is completely consistent; path 2 maps 3 technical means nodes and 1 technical constraint node, and there is 1 topological order violation. According to the scheme of this application, path 1 will be preferentially selected as the final structure mapping because it maps more constraint nodes and the number of topological order violations is zero.

[0094] Through the above technical solution, this application effectively addresses the ambiguity and inaccuracy that may arise when performing structural mapping in complex technical paths. By establishing functional anchor points, considering topological order and connection structure, and optimizing the selection of mapping paths, it ensures that the established structural mapping is semantically consistent, logically coherent, and the most complete. This significantly improves the accuracy and reliability of subsequent constraint comparisons and dominance determination results, thereby enhancing the overall effectiveness of the intellectual property monitoring and early warning system and effectively avoiding misjudgments or omissions caused by mapping errors.

[0095] In some embodiments described above in this application, a structural mapping is proposed to establish a structural mapping between the target technical path and the claim technical paths in the technical path knowledge graph that are alignable at the technical function node level, and constraints are compared point by point under the structural mapping. However, when comparing specific parameter range constraints, how to accurately determine whether the parameter range constraints of the claim technical path are not weaker than the parameter range constraints of the target technical path, especially when the parameter units are inconsistent or there are open boundaries, is a problem that needs to be solved. This directly affects the accuracy and reliability of constraint partial order determination. In this regard, this application further proposes a method for determining constraint partial order according to the parameter range inclusion relationship, specifically including: The parameter names of the corresponding parameter range constraint nodes in the target technical path and the claim technical path are standardized in terms of units and merged with those of the same name. "Unit standardization" refers to converting parameter values ​​with different units of measurement into a single standard unit, such as standardizing length to meters and temperature to degrees Celsius, to eliminate the influence of unit differences on the comparison results. This can be achieved through a pre-defined unit conversion rule base or by calling an external unit conversion service. "Merging with the same name" refers to identifying and processing parameter names that are semantically identical but have different expressions. For example, "temperature" and "temperature value" may refer to the same physical quantity. By establishing a thesaurus or ontology mapping for parameter names, these names are unified to ensure semantic consistency of the compared parameters.

[0096] Read the lower and upper limits of the constraint nodes corresponding to the parameter intervals of the claimed technical path, and read the lower and upper limits of the constraint nodes corresponding to the parameter intervals of the target technical path. This step aims to obtain the basic data required for parameter interval comparison. These lower and upper limits can be directly extracted from the constraint node attributes pre-stored in the technical path knowledge graph, or dynamically obtained by parsing structured constraint text.

[0097] When the lower limit of the parameter range constraint node corresponding to the technical path of the claim is greater than or equal to the lower limit of the parameter range constraint node corresponding to the technical path of the target technical path, and the upper limit of the parameter range constraint node corresponding to the technical path of the claim is less than or equal to the upper limit of the parameter range constraint node corresponding to the technical path of the target technical path, it is determined that the parameter range constraint of the technical path of the claim is not weaker than the parameter range constraint of the technical path of the target technical path. This determination logic is based on the interval inclusion relationship; that is, if the parameter range of the technical path of the claim is completely contained within the parameter range of the technical path of the target technical path, or the two are completely identical, then the constraint of the technical path of the claim is considered to be more stringent or at least not weaker than the constraint of the technical path of the target technical path. This can be directly implemented using conditional statements in programming languages.

[0098] When any parameter name lacks an upper or lower limit value in the target technical path, the missing boundary is marked as an open boundary, and only the corresponding boundaries of the claim technical paths are allowed to be the same open boundary in the constraint partial order determination. This step aims to handle common open interval cases in parameter range constraints, such as "temperature not lower than 20 degrees Celsius" or "pressure not exceeding 100 kPa". For missing boundaries, they are marked as open boundaries (e.g., represented by positive infinity or negative infinity). When performing constraint partial order determination, if a boundary of the target technical path is open, then partial order determination is only allowed if the corresponding boundary of the claim technical path is also open and the opening direction is consistent. For example, if the upper limit of the target technical path is open, the upper limit of the claim technical path must also be open to continue comparing the lower limit. This ensures accurate comparison of incomplete constraints and avoids misjudgment due to missing data.

[0099] This application's solution first normalizes the units of parameter names and merges names with those of the same name, ensuring that parameters from different sources and expressions have a unified benchmark before comparison, thereby eliminating comparison errors that may be caused by data inconsistencies. Subsequently, through a clear logic of interval inclusion relationships, it can accurately determine whether the parameter interval constraints of the claimed technical path are more stringent or equivalent to the constraints of the target technical path. Furthermore, for common open boundary cases in parameter intervals, this solution introduces a special handling mechanism for open boundaries, ensuring reasonable and accurate partial order determination even when constraint definitions are incomplete. This systematic approach makes the comparison process of parameter interval constraints more robust and accurate, providing a solid foundation for subsequent determination of dominance relationships.

[0100] The following is a specific example to illustrate this. Suppose that the target technical path includes a parameter constraint of "temperature between 20°C and 30°C", while the claim technical path includes multiple candidate parameter constraints.

[0101] First, all parameter names are standardized by unit and merged with identical names. For example, if "temperature value" or "degrees Celsius" appears in the technical path of the claim, it will be unified as "temperature" and "°C".

[0102] Next, the lower limit of the target technology path, 20°C, and the upper limit, 30°C, are read.

[0103] For a candidate constraint "temperature between 25°C and 28°C" in the technical path of the claim, its lower limit of 25°C and upper limit of 28°C are read. Since 25°C is greater than or equal to 20°C and 28°C is less than or equal to 30°C, it is determined that the parameter range constraint of the technical path of the claim is not weaker than the parameter range constraint of the target technical path.

[0104] Consider another candidate constraint, "temperature not lower than 22℃". In this case, the upper limit of the target technical path is 30℃, while the upper limit of the claim technical path is missing (an open boundary, which can be considered positive infinity). According to the rules, since the upper limit of the target technical path is not an open boundary, while the upper limit of the claim technical path is, the parameter range constraint of the claim technical path is determined to be weaker than the parameter range constraint of the target technical path.

[0105] If the constraint of the target technical path is "temperature not lower than 20℃", while the constraint of the claim technical path is "temperature not lower than 22℃", then both upper limits are open boundaries. In this case, since the lower limit of the claim technical path, 22℃, is greater than or equal to the lower limit of the target technical path, 20℃, and both upper limits are open boundaries, it is determined that the parameter range constraint of the claim technical path is not weaker than the parameter range constraint of the target technical path.

[0106] Through the above technical solution, this application can accurately and reliably compare the parameter range constraints in the target technical path and the claim technical path, effectively solving the comparison difficulties caused by inconsistent parameter units, differences in name expression, or the existence of open boundaries. This significantly improves the accuracy of constraint order determination, thereby enabling the intellectual property monitoring and early warning system to more accurately identify the dominance relationship between technical solutions, avoiding misjudgments or omissions caused by improper comparison of parameter constraints, and thus improving the reliability and effectiveness of the early warning results.

[0107] In some of the embodiments described above in this application, a knowledge graph-based intellectual property monitoring and early warning method is proposed. By constructing a technical path knowledge graph and establishing a structural mapping, the constraint relationships between technical solutions can be initially compared. However, in practical applications, technical solutions often contain complex execution order requirements, such as fixed order, partial order, or interchangeable order. If only simple constraint comparisons such as parameter ranges are relied upon, it may be impossible to accurately determine the strength of the relationship between different technical paths in terms of execution order, leading to incomplete or inaccurate determination of the dominance relationship, thereby affecting the accuracy of the early warning.

[0108] In response, this application further proposes a step for determining the partial order of constraints according to the strength of the execution order, including: dividing the sequence constraint nodes into fixed sequence constraints, partial sequence constraints, and commutative sequence constraints; when the sequence constraint node corresponding to the claim technical path is a fixed sequence constraint and the sequence constraint node corresponding to the target technical path is a partial or commutative sequence constraint, determining that the sequence constraint corresponding to the claim technical path is not weaker than the sequence constraint corresponding to the target technical path; when the sequence constraint node corresponding to the claim technical path is a partial sequence constraint, verifying whether all the constrained preceding technical means nodes in the target technical path are located before the subsequent technical means nodes based on the directed edge relationship after structural mapping, and determining that the constraint partial order is satisfied when the verification is successful; when a condition triggering edge appears in the structural mapping, treating the path segment corresponding to the condition triggering edge as an independent sequence domain, and performing the order strength determination separately.

[0109] The classification of sequence constraint nodes into fixed sequence constraints, partial sequence constraints, and commutative sequence constraints refers to categorizing nodes in the technology path knowledge graph that define the execution order of technical means. Fixed sequence constraints mean that technical means must be executed strictly according to a pre-defined unique sequence, without any changes. For example, step A must precede step B, and step B must precede step C, forming a unique path A->B->C. Partial sequence constraints mean that there are certain mandatory sequential relationships between technical means, but the relative order of other technical means can be flexibly adjusted. For example, step A must precede step B, but step C can be between or after A and B, as long as A precedes B. Commutative sequence constraints mean that the execution order of technical means can be arbitrarily interchanged without affecting the final result. For example, steps A and B can be executed A->B or B->A. This classification can be achieved by adding a "constraint type" field to the constraint nodes of the knowledge graph to store the identifier of "fixed sequence," "partial sequence," or "commutative sequence," or by inferring the sequence type by analyzing the connection relationships and attributes of the constraint nodes.

[0110] When the sequence constraint node corresponding to the claim's technical path is a fixed sequence constraint and the sequence constraint node corresponding to the target technical path is a partial sequence constraint or a commutative sequence constraint, the determination that the sequence constraint corresponding to the claim's technical path is not weaker than the sequence constraint corresponding to the target technical path aims to clarify the comparison rules for the strength of different types of sequence constraints. Fixed sequence constraints are the most stringent because they do not allow any changes, while partial and commutative sequences offer greater flexibility. Therefore, when the constraint of the claim path is more stringent (fixed sequence) and the constraint of the target path is more lenient (partial or commutative), the constraint of the claim path is considered not weaker than the target path. This can be achieved by setting priority rules in the comparison logic; for example, defining a sequence strength level: fixed sequence > partial sequence > commutative sequence. During the comparison, if the sequence constraint level of the claim path is higher than or equal to the sequence constraint level of the target path, it is determined to be not weaker.

[0111] When the sequence constraint nodes corresponding to the claim's technical path are partially sequence constraints, this method verifies whether all constrained preceding technical means nodes in the target technical path are located before subsequent technical means nodes based on the directed edge relationships after structural mapping. If the verification is successful, the constraint partial order is determined to be satisfied. This provides a specific verification method for partially sequence constraints. It ensures that when the claim path has partial sequence constraints, the target path must fully comply with these mandatory preceding-subsequent relationships. After structural mapping, for each partial sequence constraint in the claim path, the preceding and subsequent technical means nodes it defines are identified. Then, in the corresponding mapped structure of the target technical path, all relevant directed edges are traversed to check whether all preceding nodes are indeed located before their corresponding subsequent nodes. This can be achieved using graph traversal algorithms such as topological sorting or depth-first search.

[0112] When a condition-triggered edge appears in the structural mapping, the path segment corresponding to the condition-triggered edge is treated as an independent sequence domain, and the order strength and weakness judgment is performed separately. This aims to solve the problem of how to correctly compare order constraints when there are conditional branches in the technical path. A condition-triggered edge means that the path may choose different branches based on specific conditions, and the order constraints within these branches should be evaluated independently of other branches. When a condition-triggered edge is detected in the structural mapping (e.g., identified by the edge's attribute or type), the subsequent path segment connected by the condition-triggered edge is treated as an independent subgraph or "sequence domain." Then, the order strength and weakness judgment rules described above are applied recursively or independently to the order constraint nodes within this independent sequence domain. This ensures that the logic of conditional branches does not interfere with the order judgment of the main path or other branches.

[0113] In the aforementioned knowledge graph-based intellectual property monitoring and early warning method, to more accurately assess the dominance relationship between technical solutions, especially in scenarios involving complex execution sequences, this application further refines the mechanism for determining the partial order of constraints based on the strength of the execution order. First, the order constraint nodes in the technical path knowledge graph are classified into fixed order constraints, partial order constraints, and commutative order constraints according to their strictness. This classification provides a clear benchmark for subsequent comparisons. When the order constraint of the claim technical path is the strictest fixed order constraint, while the corresponding constraint of the target technical path is a relatively loose partial order or commutative order, it is directly determined that the order constraint of the claim technical path is not weaker than that of the target technical path, reflecting the dominance of the stricter constraint over the looser constraint. For cases where there are partial order constraints in the claim technical path, based on the established structural mapping, all preceding technical means nodes in the target technical path limited by this partial order constraint are meticulously verified to ensure that they are indeed all located before their corresponding subsequent technical means nodes in the actual execution order. This verification process is completed by analyzing the directed edge relationships after mapping, thereby ensuring the effectiveness and accuracy of the partial order constraints. Furthermore, considering the potential conditional branches in the technical solution, this application introduces the concept of independent sequence domains. When a conditional trigger edge appears in the structural mapping, the subsequent path segment guided by that trigger edge will be regarded as an independent sequence domain. The strength of the sequence constraints within these independent sequence domains will be determined separately, avoiding interference from conditional branch logic on the overall sequence judgment and ensuring the accuracy and completeness of sequence constraint comparison under complex paths.

[0114] As a specific implementation, suppose there exists a target technical path containing "step A, step B, and step C," where there is an interchangeable order constraint between step A and step B, meaning the execution order of steps A and B can be interchanged. Now, it needs to be compared with a claim technical path that also contains "step A, step B, and step C," but there is a fixed order constraint between steps A and B, meaning step A must be executed before step B. According to the solution of this application, firstly, the "interchangeable order constraint between step A and step B" in the target technical path is identified as an interchangeable order constraint node, and the "fixed order constraint between step A and step B" in the claim technical path is identified as a fixed order constraint node. When determining the constraint partial order, since the order constraint node of the claim technical path is a fixed order constraint, while the corresponding order constraint node of the target technical path is an interchangeable order constraint, according to the preset strong-weak relationship rule, the fixed order constraint is not weaker than the interchangeable order constraint. Therefore, at this mapping point, the constraint partial order of the claim technical path is satisfied. For another example, if the technical path of the claim contains a partial order constraint, such as "step X must be executed before step Y," and the target technical path also contains corresponding "step X" and "step Y" after structural mapping, then the directed edge relationship from "step X" to "step Y" in the target technical path will be checked to confirm that "step X" is indeed located before "step Y." If the check is successful, then the partial order of the partial order constraint is satisfied. Furthermore, if a condition-triggered edge is found in the technical path of the claim during structural mapping, such as "if condition M is true, then execute step P," then "step P" and its subsequent path segments will be considered an independent order domain. The strength of the order constraint within this independent order domain (e.g., the order of step P and subsequent steps) will be determined only, without confusing it with the order constraint of the path segment before the condition-triggered edge, thus ensuring the accurate evaluation of the conditional branch logic.

[0115] Through the aforementioned technical solution, this application enables refined classification and accurate comparison of complex execution order constraints within technical solutions. By dividing order constraints into fixed order, partial order, and commutative order, and formulating clear partial order determination rules for different types, it effectively solves the problem of inaccurate judgments that may occur when traditional methods handle diverse sequence logic. In particular, the meticulous verification of partial order constraints and the separate determination of independent sequence domains formed by conditional triggering edges ensure a more comprehensive and accurate assessment of the strength of execution order relationships in complex technical paths and conditional branching scenarios. This significantly improves the accuracy and reliability of intellectual property monitoring and early warning methods in identifying the dominance relationship of technical solutions, avoiding misjudgments or omissions caused by deviations in sequence logic judgments, thereby providing users with more accurate early warning results.

[0116] In some of the above-mentioned implementations, methods for determining constraint partial order based on parameter interval inclusion relationships and execution order strength relationships have been proposed. However, in actual technical solution comparisons, the implementation methods of technical means are often subject to exclusive or set-based limitations, such as "can only be implemented through method A" or "can be implemented through methods B and C, but not through method D". If these exclusive limitations are not accurately identified and compared, it may lead to inaccurate judgment of the dominance relationship of technical solutions, thereby affecting the effectiveness of intellectual property monitoring and early warning. In this regard, this application further proposes a method for determining constraint partial order based on the consistency of exclusive limitations. The steps include: parsing the exclusive constraint node in the technical path of the claim as a unique method limitation or a restricted set limitation, and writing the restricted set limitation into the set of allowed implementation methods; under structural mapping, reading the implementation method identifier of the technical means node corresponding to the target technical path; when the exclusive constraint node is a unique method limitation, verifying that the implementation method identifier corresponding to the target technical path is consistent with the unique method identifier; when the exclusive constraint node is a restricted set limitation, verifying that the implementation method identifier corresponding to the target technical path belongs to the set of allowed implementation methods; when the verification fails, determining that the constraint partial order of the mapping point is not satisfied and terminating the dominance determination.

[0117] In a technology path knowledge graph, an exclusive constraint node is a node that restricts a specific technology means node to adopt a particular implementation method, or can only adopt one of a specific set of implementation methods, or cannot adopt one of a specific set of implementation methods. It represents the exclusivity or restriction of a technology means in its implementation. An exclusive constraint node can be a Boolean type marker; when true, it indicates that the corresponding technology means node can only be implemented through a predefined, unique method. Alternatively, an exclusive constraint node can be a list or set containing identifiers of allowed or prohibited implementation methods, used to limit the scope of technology means implementation. A unique method constraint means that the exclusive constraint node explicitly specifies that the technology means node it constrains can only be completed through a specific, unique implementation method. This constraint is strongly exclusive, disallowing any other implementation methods. A unique method constraint can be represented by directly storing a string or enumeration value in the exclusive constraint node, representing a unique implementation method identifier; or it can be represented by pointing to a specific entry in a predefined implementation method dictionary or ontology, ensuring its uniqueness and standardization. A restricted set constraint refers to an exclusive constraint node specifying the set of implementation methods that the constrained technical means node can or cannot adopt. This constraint allows technical means to choose implementation methods within a certain range, but also imposes clear boundary restrictions. A restricted set constraint can be a list or array containing multiple implementation method identifiers, representing allowed or prohibited implementation methods; alternatively, it can be implemented by defining a set of rules or logical expressions used to determine whether a certain implementation method belongs to the allowed or prohibited set. The allowed implementation method set is a specific manifestation of a restricted set constraint, explicitly listing the identifiers of all implementation methods that a technical means node is allowed to adopt. When an exclusive constraint node is a restricted set constraint, these allowed implementation methods are collected to form a set for comparison. The allowed implementation method set can be a dynamically constructed hash set used to quickly find and determine whether a certain implementation method identifier exists within it; alternatively, it can be a predefined enumeration type or a list of constants directly referenced when resolving the exclusive constraint node. An implementation method identifier is a symbol or code used to uniquely identify the specific implementation method of a technical means node. It can associate abstract technical means with specific engineering implementations or technical details. The implementation method identifier can be a string, such as "mechanical transmission", "electromagnetic drive", "software control", etc., used to describe the implementation principle or type of the technical means; or it can be a numerical code or hash value, which is associated with specific implementation methods through a pre-established mapping table, facilitating internal processing and comparison.Verifying the consistency between the implementation method identifier and the unique method identifier of the target technical path aims to determine whether the implementation method of a certain technical means in the target technical path is completely identical to the unique implementation method specified by the exclusive constraint node in the claim technical path. This ensures a high degree of matching in the implementation details of the technical solutions. This can be achieved by directly determining whether the two implementation method identifiers are completely equal through string comparison or hash value comparison; alternatively, a predefined semantic similarity algorithm can be used to determine whether the two implementation method identifiers have semantic consistency within a certain tolerance range to address minor differences in description. Verifying that the implementation method identifier of the target technical path belongs to the set of allowed implementation methods is used to determine whether the implementation method of a certain technical means in the target technical path is included within the set of allowed implementation methods specified by the exclusive constraint node in the claim technical path. This ensures that the implementation method of the target technical path is within the scope of the claims. This can be achieved by searching within the set of allowed implementation methods to determine whether the implementation method identifier of the target technical path exists in that set; alternatively, it can be achieved by traversing the set of allowed implementation methods and comparing each one with the implementation method identifier of the target technical path until a match is found or the traversal ends. Determining that the partial order constraint of the mapping point is not satisfied and terminating the dominance determination means that when the exclusivity check fails, it indicates that the implementation of the target technical path does not conform to the limitations of the technical path in the claims. In this case, the partial order constraint of the mapping point is not satisfied, and since the exclusivity limit is usually critical, the dominance determination of the entire technical path can be terminated immediately to avoid unnecessary subsequent calculations. A Boolean flag can be set; if the check fails, the flag is set to false, and this flag is checked in subsequent dominance determination processes. If it is false, the determination fails directly. Alternatively, the current dominance determination process can be interrupted by throwing a specific exception or returning a specific error code, and the result of the partial order constraint not being satisfied can be reported to the upper-level module.

[0118] In the aforementioned knowledge graph-based intellectual property monitoring and early warning method, to more accurately assess the dominance relationship between the target technical path and the claim technical path, this application further introduces exclusive constraint consistency as an important basis for determining the constraint partial order. This method first performs in-depth analysis of the exclusive constraint nodes in the claim technical path, classifying them into two types: "unique mode constraint" and "restricted set constraint." For "restricted set constraint," a "set of allowed implementations" is further constructed, explicitly listing all permitted implementations. Subsequently, under the established structural mapping, the implementation method identifiers of the technical means nodes corresponding to the claim technical path in the target technical path are accurately read. Next, the method performs targeted verification based on the specific type of the exclusive constraint node. If the exclusive constraint node is a "unique mode constraint," the implementation method identifier of the target technical path is directly compared to ensure complete consistency with the unique mode identifier specified in the claim technical path. This rigorous consistency verification ensures a high degree of matching between the target technical solution and the claim in key implementation details. If the exclusive constraint node is a "restricted set constraint," the implementation method identifier of the target technical path is verified to be included in the pre-constructed "set of allowed implementations." This set-based verification mechanism allows target technical solutions to choose their implementation methods within a certain range, while ensuring that they do not exceed the boundaries defined by the claims. Through the aforementioned refined exclusive constraint consistency verification, this application can effectively identify target technical paths that conflict with the claims in terms of implementation methods. Once the exclusive constraint verification of any mapping point fails, it is determined that the constraint partial order of that mapping point is not satisfied, and subsequent dominance determination is immediately terminated. This mechanism avoids technical solutions with fundamental differences in implementation methods being mistakenly judged as having a dominance relationship, thereby significantly improving the accuracy and efficiency of intellectual property monitoring and early warning. It makes up for the shortcomings of relying solely on parameter ranges and execution order for constraint partial order determination, making the infringement risk assessment of technical solutions more comprehensive and reliable.

[0119] As a specific implementation method, special processing is performed on the consistency of exclusive constraints when determining the constraint partial order. For example, suppose there is an exclusive constraint node in the claim's technical path that restricts a certain "data transmission" technology node to be implemented via "wireless communication". In this case, the exclusive constraint node will be resolved as a "unique method restriction", and the "unique method identifier" will be set to "wireless communication". Under the structural mapping, if the implementation method identifier of the corresponding "data transmission" technology node in the target technical path is "wired communication", then it is checked whether "wired communication" and "wireless communication" are consistent. Since they are inconsistent, the constraint partial order of this mapping point will be determined to be unsatisfactory, and the determination of dominance over the target technical path will be terminated immediately. As another example, if the exclusive constraint node in the claim's technical path restricts a certain "data storage" technology node to be implemented via "cloud storage" or "local storage". In this case, the exclusive constraint node will be resolved as a "restricted set restriction", and the "allowed implementation set" will be constructed to include both "cloud storage" and "local storage" identifiers. Under the structural mapping, if the implementation method of the corresponding "data storage" technology node in the target technology path is identified as "distributed storage," it is checked whether "distributed storage" belongs to the "set of allowed implementation methods." If "distributed storage" is not in this set, the constraint partial order of this mapping point is also determined to be unsatisfactory, and the dominance determination is terminated. Conversely, if the implementation method of the target technology path is identified as "cloud storage," the verification passes, and the dominance determination continues.

[0120] Through the above technical solution, this application can perform refined comparison and limitation of the implementation methods involved in the technical solution, solving the blind spots that may exist in traditional methods when dealing with exclusive or collective implementation method constraints. This method resolves exclusive constraint nodes as unique method limitations or restricted set limitations, and constructs a set of allowed implementation methods, making the verification of the target technical path implementation methods more accurate and comprehensive. When the implementation method of the target technical path does not conform to the exclusive limitation of the claim technical path, it can promptly determine that the constraint partial order is not satisfied and terminate the dominance determination, avoiding the erroneous determination that a solution that does not conform to the technical limitation is a valid dominance relationship. This significantly improves the accuracy and reliability of intellectual property monitoring and early warning, effectively reduces the risk of misjudgment, and makes the early warning results more instructive, thereby providing users with a more accurate infringement risk assessment.

[0121] In knowledge graph-based intellectual property monitoring and early warning methods, accurately determining whether a target technical path is "occupied" by a claim technical path requires precisely defining the boundaries of the constraints within the claim technical path. However, in practice, if the definition of the scope of the constraints is not clear enough, it may lead to ambiguity or inaccuracy in determining the occupancy status of the target technical path, thereby affecting the reliability of intellectual property monitoring and early warning.

[0122] In response, this application further proposes a method for determining whether a technical means node of a target technical path falls within the scope of constraint of the technical path of the claims. This method includes: recording the starting point technical means node identifier and the ending point technical means node identifier for each technical constraint node; determining a set of reachable path segments from the starting point technical means node identifier to the ending point technical means node identifier based on directed connection relationships, and defining the set of reachable path segments as the scope of constraint; mapping each technical means node of the target technical path to a corresponding technical means node of the technical path of the claims under structural mapping, and verifying whether the corresponding technical means node belongs to the scope of constraint; when all technical means nodes of the target technical path satisfy the verification, determining that the path occupancy status is valid.

[0123] This process involves recording the starting and ending technical means node identifiers for each technical constraint node, aiming to clarify the effective and ineffective boundaries of the constraint imposed by each node. Technical constraint nodes may impose limitations on the function, performance, or implementation method of technical means nodes. By recording the starting and ending technical means node identifiers, the scope of influence of the constraint can be precisely defined. For example, in one implementation, two attribute fields can be added to each technical constraint node in the technical path knowledge graph to store the starting and ending technical means node IDs of its scope of influence, respectively. In another implementation, these starting and ending identifiers can be automatically identified and labeled by analyzing the syntactic structure and semantic relationships of constraint phrases in the claim text during knowledge graph construction. Based on directed connections, a set of reachable path segments from the starting and ending technical means node identifiers is determined, and this set of reachable path segments is defined as the constraint scope. This transforms discrete starting and ending identifiers into a continuous, directed set of path segments, thus forming a clearly defined constraint area. This area contains all technical means nodes affected by the constraint. For example, graph traversal algorithms (such as depth-first search or breadth-first search) can be used to explore all reachable paths along directed connections, starting from the starting technical means node identifier, until the ending technical means node identifier is encountered, and all technical means nodes on these paths are collected. Another implementation is to pre-compute and store a list of all affected technical means nodes corresponding to each technical constraint node in the knowledge graph to improve query efficiency. Under structural mapping, mapping each technical means node in the target technical path to the corresponding technical means node in the claim technical path and verifying whether the corresponding technical means node belongs to the scope of constraint is the core step in actual judgment. It compares the specific technical means in the target technical solution with the scope of constraint defined in the patent claim. Through structural mapping, the technical means node in the target technical path finds its corresponding position in the claim technical path. Then, it checks whether the technical means node at this corresponding position falls within the previously defined scope of constraint. For example, all technical means nodes in the target technical path can be traversed, and their corresponding nodes in the claim technical path can be found through the established structural mapping. Then, it is queried whether the corresponding node is in the set of constraint scope. Another approach is to directly mark the mapped target technical means nodes after the structural mapping is established, indicating whether they are covered by a constraint in the technical path of the claim. When all technical means nodes of the target technical path meet the verification, the path occupancy status is determined to be valid, which is the final condition for determining the "path occupancy status".Only when all technical means nodes in the target technical path are covered by the constraint scope of the claim technical path after being mapped to the claim technical path, is the target technical path considered to have completely "occupied" the constraint space of the claim technical path. For example, a Boolean flag can be set, initially true. During each verification process, if any technical means node is found to fail the verification, the flag is set to false. After all verifications are completed, the path occupancy status is determined based on the final state of the flag.

[0124] This application defines a clear effective area for each constraint in the technology path knowledge graph by explicitly recording the starting and ending technical means node identifiers of the constraint's scope of action for each technical constraint node. Subsequently, based on the directed connections already constructed in the knowledge graph, starting from the starting technical means node identifier and following the logical flow of the technology path to the ending technical means node identifier, the set of all technical means nodes covered by the constraint—that is, the constraint's scope of action—is dynamically determined and defined. After completing the structural mapping between the target technology path and the claim technology path, this application locates each technical means node in the target technology path to its corresponding technical means node in the claim technology path through this mapping relationship. Then, these mapped technical means nodes are verified one by one to determine whether they are all included within the previously defined constraint scope. This mechanism ensures that the path occupancy status is determined only when all technical means nodes in the target technology path are under the constraint of the claim technology path. In this way, this application avoids misjudgment caused by unclear scope of constraint, making the judgment of the occupancy status of the target technical path more accurate and reliable, thereby improving the accuracy of intellectual property monitoring and early warning.

[0125] As a specific implementation, suppose there exists a technical constraint node in the technical path of the claims, which limits the "processing speed" of the "processing unit" to be between "100MHz and 200MHz". To clarify the scope of this speed constraint, the starting technical means node of this technical constraint node can be identified as the "data receiving module" and the ending technical means node as the "data sending module". Based on the directed connection relationship in the technical path knowledge graph, starting from the identifier of the "data receiving module", all reachable technical means nodes are traversed along the direction of data flow or control flow until the identifier of the "data sending module". For example, this path may sequentially pass through the "data receiving module", "data preprocessing module", "processing unit", "data postprocessing module" and "data sending module". Then, the set of path segments composed of these modules is defined as the scope of the "processing speed" constraint. When a target technical path is mapped to the technical path of the claims, for example, the "target data processing module" in the target technical path is mapped to the "processing unit" in the technical path of the claims, it is then verified whether the "processing unit" belongs to the previously determined scope of the constraint. If the "processing unit" is indeed within this scope, and all other technical means nodes in the target technical path are also covered by the corresponding constraint scope of the claimed technical path after mapping, then the path occupancy status can be determined to be valid.

[0126] Through the above technical solution, this application can accurately define the boundaries of the influence of each technical constraint node in the technical path of the claims, avoiding the ambiguity in judgment caused by unclear constraint scope. By recording clear start and end point technical means node identifiers for each technical constraint node, and dynamically determining its reachable path segment set as the constraint scope based on directed connection relationships, the judgment of whether technical means nodes in the target technical path fall within the constraint scope of the technical path of the claims has high accuracy and traceability. This significantly improves the accuracy of path occupancy status determination, thereby effectively reducing false alarms and missed alarms in the intellectual property monitoring and early warning process, making the early warning results more reliable, and providing users with a more accurate intellectual property risk assessment.

[0127] In some of the aforementioned implementations, while it is possible to determine the dominance relationship between the claim's technical path and the target technical path through structural mapping and constraint partial order comparison, and further determine whether the technical means nodes of the target technical path fall within the constraint scope of the claim's technical path, thereby marking the path occupancy status, it may be difficult to accurately identify the core technical elements in the claim's technical path that constitute substantial and difficult-to-avoid obstacles to the target technical solution based solely on the dominance relationship and path occupancy status. This results in insufficient targeting of the early warning results.

[0128] In response, this application further proposes steps for identifying unavoidable structural nodes and outputting structural conflict reasoning results, specifically including: determining technical means nodes in the claim technical path with at least 2 preceding directed edges entering the same technical means node as branch-merging nodes; constructing a set of feasible directed paths with the technical function nodes of the claim technical path as the starting point and the technical function completion nodes as the ending point; when any feasible directed path contains the same branch-merging node, determining the branch-merging node as a necessary node; when there are exclusive constraint nodes or dependent constraint nodes among the constraint nodes associated with the necessary node, marking the necessary node as an unavoidable structural node; the structural conflict reasoning result includes at least a dominance establishment marker, a path occupancy status marker, a set of unavoidable structural node identifiers, the mapping path segment identifier corresponding to the unavoidable structural node, and the constraint node identifier that does not satisfy the constraint partial order.

[0129] In this context, a branch-convergence node refers to a situation in the technical path of a claim where two or more different preceding technical means nodes are connected to the same technical means node through directed edges. This technical means node, as the intersection of multiple technical paths, represents the convergence of different implementation paths in the technical solution. Its function is to identify key integration points or shared resource points in the technical process. For example, by traversing the technical path knowledge graph and counting the in-degree (i.e., the number of directed edges pointing to that node) of each technical means node, if the in-degree is greater than or equal to 2, it is marked as a branch-convergence node; alternatively, when constructing the knowledge graph, a list of preceding nodes can be maintained for each technical means node, and when the number of nodes in the list reaches or exceeds two, it is determined to be a branch-convergence node.

[0130] A necessary node refers to a specific branch-and-merge node in the technical path of a claim, which must be traversed by all possible technical implementation paths from a technical function node to a technical function completion node. This feature aims to identify key, irreplaceable links in the technical solution from numerous branch-and-merge nodes. For example, depth-first search (DFS) or breadth-first search (BFS) algorithms can be used to explore all directed paths to the technical function completion node, starting from the technical function node, and recording the branch-and-merge nodes traversed by each path. If a branch-and-merge node appears in all these feasible paths, it is determined to be a necessary node; alternatively, the concepts of cut vertices or bridges in graph theory can be used to determine necessary nodes by analyzing whether removing a branch-and-merge node would cause the technical function node and the technical function completion node to become disconnected.

[0131] Unavoidable structural nodes refer to those nodes that are not only essential steps in a technical solution but also subject to strong constraints (such as exclusive or strong dependency constraints). This feature further focuses on core technical points that are difficult to circumvent through alternative solutions. For example, after identifying an unavoidable node, query the constraint nodes that are directly or indirectly related to it. If these constraint nodes contain exclusive constraint nodes (e.g., limiting the use to a specific technical implementation) or dependency constraint nodes (e.g., forcing a binding to a specific technical means or condition), then the unavoidable structural node is marked as an unavoidable structural node; alternatively, a list of associated constraint types can be maintained for each unavoidable node, and when the list contains "exclusive" or "dependency" types, it is marked as an unavoidable structural node.

[0132] The structural conflict reasoning result is a comprehensive analysis report, the final output of this method. It not only includes the dominance establishment markers and path occupancy status markers determined in previous steps, but also further refines the nature and location of the conflict. This result aims to provide users with comprehensive and instructive information on intellectual property risks. For example, the result can be presented in a structured data format (such as a JSON object or XML document), containing dominance establishment markers and path occupancy status markers represented by Boolean values, a list or set storing unique identifiers for all unavoidable structural nodes, another list or set storing the specific mapping path segment identifiers of these unavoidable structural nodes in the claim's technical path, and a list or set recording the identifiers of specific constraint nodes that failed to meet the conditions in the constraint partial order comparison.

[0133] This method, based on determining the dominance relationship between the claim's technical path and the target technical path, as well as the path occupancy status, further deepens the identification of potential conflict points. First, by analyzing the connection relationships of technical means nodes in the claim's technical path, it identifies branch convergence nodes that serve as convergence points for multiple technical processes. These nodes represent the intersection of different implementation paths in the technical solution. Subsequently, this method constructs a set of all feasible directed paths from technical function nodes to technical function completion nodes, precisely filtering out those branch convergence nodes that all paths must traverse, classifying them as essential nodes. This process ensures that the identified nodes are indispensable key links in the technical solution. Based on this, the method further examines the constraint nodes associated with these essential nodes. If exclusive constraint nodes or dependent constraint nodes are found, the essential node is marked as an unavoidable structural node. This means that the node is not only an essential path in the technical solution, but its implementation method or conditions are also subject to strong and unavoidable constraints. Finally, this method integrates these identified unavoidable structural nodes, along with dominance establishment markers, path occupancy status markers, mapping path segment identifiers corresponding to unavoidable structural nodes, and constraint node identifiers that do not satisfy the constraint partial order, into a structural conflict reasoning result. Through this progressive analysis, this method can delve from macroscopic dominance relationships to microscopic, unavoidable technical details, providing users with a comprehensive, accurate, and highly instructive intellectual property conflict early warning, significantly improving the depth and practicality of the warning.

[0134] As a specific implementation, suppose there exists a technical path described in the claims, outlining a "smart home control system," where the technical function node is "user command reception" and the technical function completion node is "device status feedback." In this path, multiple input modules (such as a voice recognition module, a touch panel module, and a mobile app module) may send processed commands to a "central processing unit" for parsing. In this case, the "central processing unit" is determined as a branch-merging node because it receives command streams from multiple preceding modules. Furthermore, if all feasible paths from "user command reception" to "device status feedback" must pass through this "central processing unit," then the "central processing unit" is determined as a necessary node. Based on this, if among the constraint nodes associated with this "central processing unit," there is an exclusive constraint node that requires the central processing unit to use a "specific type of embedded processor," or a dependency constraint node that requires it to run a "self-developed operating system kernel," then this "central processing unit" will be marked as an unavoidable structural node. Ultimately, the output structural conflict reasoning result will explicitly indicate the dominance establishment marker (if applicable), the path occupancy status marker (if applicable), and specifically list the "Central Processing Unit" as an unavoidable structural node. It will also provide the specific mapping path segment identifier in the technical path of the claims, as well as the constraint node identifiers such as "specific model of embedded processor" or "self-developed operating system kernel" that make it unavoidable.

[0135] Through the above technical solution, this method overcomes the limitations of relying solely on dominance relationships and path occupancy status for early warning, achieving refined identification of intellectual property conflict points. By introducing the concepts of branch convergence nodes, mandatory nodes, and unavoidable structural nodes, and combining them with the limitations of exclusive constraint nodes or dependency constraint nodes, this method can accurately locate the core technical elements in the claim's technical path that constitute substantial and difficult-to-avoid obstacles to the target technical solution. This makes the intellectual property monitoring and early warning results more targeted and instructive, helping users clearly understand the nature and location of potential conflicts, thereby effectively supporting risk avoidance, avoidance design, and innovation decisions for technical solutions, and avoiding blind investment or unnecessary legal risks.

[0136] In another preferred embodiment based on the above embodiments, see [reference] Figure 2 As shown, this embodiment provides an intellectual property monitoring and early warning system based on knowledge graphs. The system includes a target path construction module, a patent set determination module, a graph construction module, a structure mapping module, a dominance relationship determination module, a path occupancy determination module, an unavoidable structure identification module, and a result output module.

[0137] The target technical path construction module is configured to acquire a technical solution description input by the user, and then analyze the technical function, technical means, and technical constraints of the technical solution description to construct the target technical path. This module can employ Natural Language Processing (NLP) techniques, such as lexical analysis, syntactic analysis, and semantic analysis, to identify key information in the text and transform it into a structured data representation. For example, it can utilize pre-trained language models or rule-based pattern matching to extract technical functions, technical means, and technical constraints. Furthermore, this module can also directly construct the target technical path by providing a structured input interface that guides the user to input key elements of the technical solution in a predefined format.

[0138] The patent set determination module is configured to identify a pre-defined set of patents in the patent data source that have a mapping relationship with the technical function nodes in the target technical path. This module can use methods such as keyword matching, semantic similarity calculation, or patent classification number (e.g., IPC, CPC) mapping to associate the technical function nodes of the target technical path with the set of technical topic nodes or classification nodes in the patent data source. For example, by constructing a technical topic dictionary or classification system, the functional descriptions in the target technical path can be mapped to corresponding technical topics or classification tags, thereby filtering relevant patents from the patent database. Alternatively, this module can use machine learning models, such as text classifiers, to classify the functional descriptions of the target technical path and retrieve patents belonging to the same or related categories from the patent data source based on the classification results.

[0139] The knowledge graph construction module is configured to acquire the claim text of a predefined patent set, parse it to obtain technical functional elements, technical means elements, and technical constraint elements, and construct a technical path knowledge graph containing technical functional nodes, technical means nodes, technical constraint nodes, and their directed connections. This module can employ information extraction techniques, such as rule-based extraction systems or deep learning models (e.g., sequence labeling models), to identify and extract technical functional elements, technical means elements, and technical constraint elements from the claim text. The extracted elements are then transformed into nodes in the knowledge graph, and directed connections are established between nodes based on the logical relationships and syntactic structures in the claim text. For example, graph databases (e.g., Neo4j, ArangoDB) can be used to store and manage the constructed technical path knowledge graph.

[0140] The structure mapping module is configured to use technical function nodes as anchors to establish a structure mapping between the target technical path and the claim technical paths in the technical path knowledge graph that are alignable at the technical function node level, maintaining directed connections and consistent constraint node positions. This module can employ graph matching algorithms, such as subgraph isomorphism algorithms or approximate graph matching algorithms, to find the corresponding structure of the target technical path in the technical path knowledge graph. During mapping, the module prioritizes ensuring the alignment of technical function nodes, and then, based on this, gradually establishes mapping relationships between technical means nodes and technical constraint nodes according to the topological order of directed edges and the consistency of technical constraint node positions. For example, a heuristic search algorithm, combined with preset matching rules and a scoring mechanism, can be used to evaluate the merits of different mapping schemes and select the optimal structure mapping.

[0141] The dominance relationship determination module is configured to compare constraints point-by-point under structural mapping. It determines the constraint partial order based on parameter range inclusion relationships, execution order strength, and consistency of exclusive constraints. When the claim technical path satisfies the constraint partial order at all mapping points, it is determined that the claim technical path dominates the target technical path. This module can implement a rule engine to execute different types of constraint comparison logic. For example, for parameter range constraints, this module compares the upper and lower limits of corresponding parameters in the claim technical path and the target technical path to determine if an inclusion relationship exists. For execution order constraints, this module analyzes the directed connection relationships between nodes to determine the strength of the order. For exclusive constraint constraints, this module verifies the consistency of the implementation. When the constraint partial order is satisfied at all mapping points, the module outputs a determination result indicating that dominance is established.

[0142] The path occupancy determination module is configured to, when a dominance relationship is established, determine whether all technical means nodes of the target technical path fall within the constraint scope of the claim technical path, and mark the path occupancy status when the determination is successful. This module first determines the effective scope of each constraint based on the start and end point identifiers of the technical constraint nodes using a graph traversal algorithm (such as depth-first search or breadth-first search). Subsequently, under the established structural mapping, the module checks each technical means node in the target technical path one by one, determining whether its corresponding node in the claim technical path is within the scope of the corresponding constraint. If all technical means nodes meet this condition, the path occupancy status is marked as successful.

[0143] The unavoidable structure identification module is configured to identify, within the technical path of the claims, nodes that simultaneously satisfy the conditions of branch convergence, are necessary for all feasible paths to the technical function completion node, and are subject to exclusive or strong dependency constraints, and mark these nodes as unavoidable structure nodes. This module can utilize connectivity analysis and pathfinding algorithms from graph theory. First, branch convergence nodes are identified by analyzing nodes with an in-degree greater than 1. Next, by traversing all feasible paths from the technical function node to the technical function completion node, nodes that must be traversed by all paths are identified, i.e., necessary nodes. Finally, these necessary nodes are checked for exclusive or strong dependency constraints, thus marking them as unavoidable structure nodes.

[0144] The output module is configured to output structural conflict reasoning results based on path occupancy status and unavoidable structural nodes. This module is responsible for integrating and visualizing the various judgment results generated during the analysis process. For example, it can generate a detailed report that includes dominance assertion markers, path occupancy status markers, a set of unavoidable structural node identifiers, the mapped path segment identifiers corresponding to the unavoidable structural nodes, and the identifiers of constraint nodes that do not satisfy the constraint partial order. This module can also provide an interactive interface, allowing users to view the detailed process and results of the knowledge graph, structural mapping, and conflict reasoning, thereby assisting users in decision-making.

[0145] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A knowledge graph-based method for monitoring and early warning of intellectual property rights, characterized in that, include: Obtain the user's input description of the technical solution and construct the target technology path; Determine the preset patent set based on the technical functional nodes in the target technology path; Obtain the claim text of a pre-defined patent set, parse it to obtain technical function elements, technical means elements and technical constraint elements, and construct a technical path knowledge graph that includes technical function nodes, technical means nodes, technical constraint nodes, technical function completion nodes and directed connection relationships; For the target technology path and the claim technology paths in the technology path knowledge graph that can be aligned at the level of technology function nodes, a structural mapping is established; the structural mapping maintains the directed connection relationship unchanged and keeps the position of the constraint nodes in the claim technology paths consistent; Under structural mapping, constraints are compared point by point. The constraint partial order is determined according to the parameter interval inclusion relationship, the strong and weak relationship of the execution order and the consistency of the exclusive limitation. When the technical path of the claim satisfies the constraint partial order at all mapping points, it is determined that the technical path of the claim dominates the target technical path. When dominance is established, it is determined whether all technical means nodes of the target technical path fall within the scope of the constraints of the technical path of the claims, and the path occupancy status is marked when it is established; In the technical path of the claims, identify all feasible paths that simultaneously satisfy branch convergence and reach the node where the technical function is completed, and are subject to exclusive constraints or strong dependency constraints, and mark unavoidable structural nodes. The output structure conflict reasoning results are based on path occupancy status and unavoidable structural nodes.

2. The knowledge graph-based intellectual property monitoring and early warning method according to claim 1, characterized in that, The pre-defined patent set, determined based on the technical functional nodes in the target technology path, includes: Map the technical function nodes in the target technology path to the preset set of technical topic nodes and category nodes in the technology path knowledge graph; Based on the set of technical topic nodes and the set of classification nodes, patents whose publication date is earlier than the generation time of the target technical path and whose technical topic nodes or classification nodes satisfy the mapping relationship are selected from the patent data source to form a preset patent set. After the pre-set patent set is formed, a patent identifier and a claim identifier are established for each patent, and the claim identifier is used as an index key for the technical path of the claim.

3. The knowledge graph-based intellectual property monitoring and early warning method according to claim 2, characterized in that, The text of the claims in the pre-defined patent set includes: Sentence segmentation is performed on the claim text and limiting phrases are identified to extract functional description fragments representing the technical purpose as technical functional elements; Identify combinations of action phrases and noun phrases that represent structures, steps, or modules as technical means elements; Identify parameter ranges, sequence words, dependency trigger words, and exclusive qualifiers used to limit the applicable conditions of technical means as technical constraint elements; The technical functional elements, technical means elements, and technical constraint elements are written into the technical path knowledge graph in a three-element structure of functional nodes, means nodes, and constraint nodes, and a constraint type field and a scope field are written for each constraint node.

4. The knowledge graph-based intellectual property monitoring and early warning method according to claim 3, characterized in that, Establishing a structure mapping that maintains consistency between directed connectivity and the location of constraint nodes includes: Establish a functional anchor mapping between the technical functional nodes of the target technical path and the technical functional nodes of the claim technical path; Under the functional anchor point mapping constraint, the technical means nodes of the target technical path are mapped to the technical means nodes in the claim technical path that have the same pre-order connection structure and subsequent order connection structure according to the directed edge topology order. Map the technical constraint nodes located between adjacent technical means nodes in the target technical path to the technical constraint nodes located between corresponding technical means nodes in the claim technical path; When there are more than two candidate mapping paths, the structure mapping that maximizes the number of constraint node mappings and has zero violations of topological order is selected as the final structure mapping.

5. The knowledge graph-based intellectual property monitoring and early warning method according to claim 4, characterized in that, Determining the constraint partial order based on the inclusion relationship of parameter intervals includes: The parameter names of the corresponding parameter range constraint nodes in the target technical path and the claim technical path are normalized and merged with those of the same name; Read the lower and upper limits of the parameter range constraint nodes corresponding to the technical path of the claim, and read the lower and upper limits of the parameter range constraint nodes corresponding to the target technical path; When the lower limit of the parameter interval constraint node corresponding to the technical path of the claim is greater than or equal to the lower limit of the parameter interval constraint node corresponding to the technical path of the target, and the upper limit of the parameter interval constraint node corresponding to the technical path of the claim is less than or equal to the upper limit of the parameter interval constraint node corresponding to the technical path of the target, it is determined that the parameter interval constraint corresponding to the technical path of the claim is not weaker than the parameter interval constraint corresponding to the technical path of the target. When any parameter name is missing an upper or lower limit value in the target technical path, the missing boundary is recorded as an open boundary, and in the constraint partial order determination, only the boundary corresponding to the claim technical path is allowed to be the same open boundary.

6. The knowledge graph-based intellectual property monitoring and early warning method according to claim 5, characterized in that, Determining the partial order of constraints based on the strength of execution order includes: Sequence constraint nodes are divided into fixed sequence constraints, partial sequence constraints, and commutative sequence constraints. When the sequential constraint node corresponding to the technical path of the claim is a fixed sequential constraint and the sequential constraint node corresponding to the technical path of the target is a partial sequential constraint or an interchangeable sequential constraint, it is determined that the sequential constraint corresponding to the technical path of the claim is not weaker than the sequential constraint corresponding to the technical path of the target. When the sequential constraint node corresponding to the technical path of the claim is a partial sequential constraint, the constraint of the preceding technical means node in the target technical path is checked based on the directed edge relationship after structural mapping. If the check is successful, the constraint partial order is determined to be satisfied. When a condition-triggered edge appears in the structure mapping, the path segment corresponding to the condition-triggered edge is treated as an independent sequence field, and the order strength relationship is determined separately.

7. The knowledge graph-based intellectual property monitoring and early warning method according to claim 6, characterized in that, Determining the constraint partial order based on the consistency of exclusive constraints includes: The exclusive constraint node in the technical path of the claim is parsed as a unique mode limitation or a restricted set limitation, and the restricted set limitation is written into the set of allowed implementation modes; Under the structural mapping, read the implementation method identifier of the technical means node corresponding to the target technical path; When the exclusive constraint node is limited by a unique method, the verification target technical path's corresponding implementation method identifier is consistent with the unique method identifier; When the exclusive constraint node is limited by a restricted set, the verification target technical path corresponds to the implementation method identifier that belongs to the set of allowed implementation methods; If the verification fails, it is determined that the partial order constraint of the mapping point is not satisfied and the dominance determination is terminated.

8. The knowledge graph-based intellectual property monitoring and early warning method according to claim 7, characterized in that, Determining whether the technical means nodes of the target technical path fall within the scope of the constraints of the technical path in the claims includes: Record the starting point technical means node identifier and the ending point technical means node identifier of the constraint's effective range for each technical constraint node; Based on the directed connection relationship, determine the set of reachable path segments from the starting technical means node identifier to the ending technical means node identifier, and define the set of reachable path segments as the scope of constraint; Under structural mapping, each technical means node of the target technical path is mapped to the corresponding technical means node of the claim technical path, and the corresponding technical means node is verified to be within the scope of the constraint. When all technical means nodes of the target technical path meet the verification requirements, the path occupancy status is determined to be valid.

9. The knowledge graph-based intellectual property monitoring and early warning method according to claim 8, characterized in that, The inference results for identifying unavoidable structural nodes and output structure conflicts include: A technical means node in the technical path of the claim with a number of preceding directed edges not less than 2 is determined as a branch-merging node; A set of feasible directed paths is constructed with the technical function nodes of the technical path of the claim as the starting point and the technical function completion nodes as the ending point. When any feasible directed path contains the same branch merging node, the branch merging node is determined to be a necessary node. When there are exclusive constraint nodes or dependent constraint nodes among the constraint nodes associated with the necessary node, the necessary node is marked as an unavoidable structural node. The structural conflict reasoning results include at least the dominance establishment marker, the path occupancy status marker, the set of unavoidable structural node identifiers, the mapping path segment identifiers corresponding to the unavoidable structural nodes, and the constraint node identifiers that do not satisfy the constraint partial order.

10. A knowledge graph-based intellectual property monitoring and early warning system, used to implement the knowledge graph-based intellectual property monitoring and early warning method according to any one of claims 1-9, characterized in that, include: The target path construction module is configured to obtain the technical solution description input by the user, and to analyze the technical function, technical means and technical constraints of the technical solution description to construct the target technical path; The patent set determination module is configured to determine a preset patent set that has a mapping relationship with the technical function nodes in the target technical path from the patent data source. The knowledge graph construction module is configured to obtain the claim text of a preset set of patents, parse out the technical function elements, technical means elements and technical constraint elements, and construct a technical path knowledge graph containing technical function nodes, technical means nodes, technical constraint nodes and their directed connection relationships. The structural mapping module is configured to use technical function nodes as anchors to establish a structural mapping that maintains directed connection relationships and consistent with the position of constraint nodes for the target technical path and the claim technical paths that can be aligned at the technical function node level in the technical path knowledge graph. The dominance relationship determination module is configured to compare constraints point by point under the structural mapping, determine the constraint partial order according to the parameter interval inclusion relationship, the strong and weak relationship of the execution order and the consistency of the exclusive limitation, and determine that the claim technical path dominates the target technical path when the claim technical path satisfies the constraint partial order at all mapping points. The path occupancy determination module is configured to determine whether all technical means nodes of the target technical path fall within the scope of the constraints of the technical path of the claims when the dominance relationship is established, and to mark the path occupancy status when the determination is established. The unavoidable structure identification module is configured to identify, in the technical path of the claim, all feasible paths that simultaneously satisfy branch convergence and reach the node of technical function completion, and are subject to exclusive constraints or strong dependency constraints, and mark the unavoidable structure nodes. The output module is configured to output structural conflict inference results based on path occupancy status and unavoidable structural nodes.

Images