Semantic feature extraction method and apparatus, device, and communication system

Abstract

A semantic feature extraction method and apparatus, a device, and a communication system. The method comprises: on the basis of a local available resource, determining to perform semantic feature extraction locally or in a cloud (S202); if it is determined to perform semantic feature extraction locally, extracting a semantic feature on the basis of a local semantic knowledge base (S204); and if it is determined to perform semantic feature extraction in the cloud, sending instruction information to a server, the instruction information being used for instructing the server to extract the semantic feature on the basis of a public semantic knowledge base (S206).

Description

Semantic feature extraction methods, devices, equipment and communication systems

[0001] This application claims priority to Chinese Patent Application No. 2024114092905, filed on October 10, 2024, entitled "Method, Apparatus, Device and Communication System for Semantic Feature Extraction", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of wireless communication technology, and in particular to a semantic feature extraction method, apparatus, device and communication system. Background Technology

[0003] Currently, semantic communication, as a novel communication paradigm, has solved the problem of how sent symbols convey precise meaning. However, faced with the surge in communication data volume, traditional solutions suffer from inefficiency in extracting semantic features. Summary of the Invention

[0004] This application provides a semantic feature extraction method, apparatus, device, and communication system that can efficiently complete semantic feature extraction.

[0005] Firstly, this application provides a semantic feature extraction method applied to a sending end, the method comprising:

[0006] Based on available local resources, determine whether to perform semantic feature extraction locally or in the cloud;

[0007] If it is determined that semantic feature extraction will be performed locally, then semantic features will be extracted based on the local semantic knowledge base;

[0008] If it is determined that semantic feature extraction will be performed in the cloud, an instruction message is sent to the server; the instruction message is used to instruct the server to extract semantic features based on a public semantic knowledge base.

[0009] In one embodiment, determining whether to perform semantic feature extraction locally or in the cloud, based on available local resources, includes:

[0010] Based on the available local resources and the resources required for semantic feature extraction, confirm whether the local system has the capability to complete semantic feature extraction.

[0011] If the local machine has the capability to perform semantic feature extraction, then compare the performance metrics of performing semantic feature extraction locally with those of performing it on the cloud, and determine whether to perform semantic feature extraction locally or on the cloud based on the comparison results.

[0012] In one embodiment, if it is determined that semantic feature extraction will be performed locally, then semantic features are extracted based on the local semantic knowledge base, including:

[0013] Semantic features are extracted based on a local semantic knowledge base;

[0014] Receive optimization suggestions from the server and optimize semantic features based on the suggestions.

[0015] In one embodiment, before the step of receiving optimization suggestions from the server and optimizing the semantic features based on the optimization suggestions, the method further includes:

[0016] Based on semantic communication goals and semantic features, prompt words are created;

[0017] The prompt words and necessary data are transmitted to the server, which instructs the server to generate optimization suggestions based on the public semantic knowledge base, and then transmit the optimization suggestions to the sender.

[0018] In one embodiment, the instruction information includes prompts and necessary data.

[0019] In one embodiment, if it is determined that semantic feature extraction will be performed in the cloud, an instruction message is sent to the server, including:

[0020] Based on semantic communication objectives and a local semantic knowledge base, create prompt words;

[0021] The prompt words and necessary data are transmitted to the server, instructing the server to extract semantic features based on a public semantic knowledge base, and then transmit the semantic features to the sender.

[0022] In one embodiment, determining whether the local system has the capability to complete semantic feature extraction based on available local resources and the resources required for semantic feature extraction includes:

[0023] The data to be processed is assessed, and the data volume information of the data to be processed is obtained;

[0024] Based on the data volume information, the resource requirements for running the local semantic feature extraction algorithm are assessed, and the resource requirements are taken as the resources required for semantic feature extraction.

[0025] If the available local resources are greater than the resources required for semantic feature extraction, then the local system is determined to have the capability to complete semantic feature extraction.

[0026] In one embodiment, the data volume information includes at least one of data size and data type; the resource requirements include at least one of computing requirements and memory requirements.

[0027] In one embodiment, the resources required for semantic feature extraction include semantic feature extraction computational requirements and semantic feature extraction memory requirements;

[0028] If the available local resources exceed the resources required for semantic feature extraction, then the local system is determined to have the capability to complete semantic feature extraction, including:

[0029] Obtain locally available resources; locally available resources include locally available computing resources and locally available memory space;

[0030] If the available local computing resources are greater than the computational requirements for semantic feature extraction, and the available local memory space is greater than the memory requirements for semantic feature extraction, then it is determined that the local system has the capability to complete semantic feature extraction.

[0031] In one embodiment, the performance metrics for semantic feature extraction performed locally include local completion time and local completion energy consumption; the performance metrics for semantic feature extraction performed offloaded to the cloud include cloud completion time, data transmission time and data transmission energy consumption for transmitting the data to be processed to the server.

[0032] Compare the performance metrics of semantic feature extraction performed locally with those performed on-premises in the cloud, and based on the comparison results, determine whether to perform semantic feature extraction locally or in the cloud, including:

[0033] The data volume information is transmitted to the server; the data volume information is used to instruct the server to calculate and report the cloud completion time, data transmission time, and data transmission energy consumption.

[0034] If the local completion time is less than the sum of the data transmission time and the cloud completion time, and the local completion energy consumption is less than the data transmission energy consumption, then semantic feature extraction will be performed locally.

[0035] If the local completion time is greater than or equal to the sum of the data transmission time and the cloud completion time, and / or the local completion energy consumption is greater than or equal to the data transmission energy consumption, then semantic feature extraction will be performed in the cloud.

[0036] In one embodiment, the local semantic knowledge base includes one or more of user information, device information, target domain knowledge, and contextual information.

[0037] In one embodiment, user information includes one or more of user profiles, user preferences, user behavior patterns, and historical interaction records; device information includes one or more of hardware configuration, software configuration, performance parameters, and device operating status; target domain knowledge includes one or more of concept definitions, entity relationships, rule logic, and domain models; and context information includes one or more of environmental state, situational information, and time information.

[0038] In one embodiment, the public semantic knowledge base includes one or more of general knowledge, domain expertise, device interaction knowledge, and a model library.

[0039] In one embodiment, general knowledge includes one or more of common sense, conceptual definitions, and general rules; domain expertise includes one or more of technical terms, cases, regulations, and industry standards; device interaction knowledge includes one or more of device types, interfaces, communication protocols, and operating instructions; and the model library includes one or more of statistical models, machine learning models, and deep learning models.

[0040] In one embodiment, semantic features include one or more of entity recognition, sentiment analysis, topic classification, and semantic role labeling.

[0041] In one embodiment, optimization suggestions include one or more of maintaining, expanding, reducing, merging, splitting, refining, and simplifying.

[0042] In one embodiment, the necessary data includes one or more of the following: semantic features, raw data, data type, data format, data creation time, and data version.

[0043] In one embodiment, the prompt word includes one or more of a communication goal description, a semantic feature summary, and the sender's intent.

[0044] In one embodiment, the prompts are presented in one or more of the following formats: text, vector, and structured data.

[0045] In one embodiment, the necessary data includes at least one of the following: raw data, data type, data format, data creation time, and data version.

[0046] In one embodiment, the prompt word includes one or more of the following: a description of the communication target, a description of the communication scenario, the sender's intent, and a personalized tag.

[0047] In one embodiment, the prompts are presented in one or more of the following formats: text, vector, and structured data.

[0048] Secondly, this application provides a semantic feature extraction method applied to a server, the method comprising:

[0049] Receive indication information from the sender;

[0050] Based on the instructions, semantic features are extracted from a public semantic knowledge base.

[0051] In one embodiment, the indication information is sent by the sending end when it determines that semantic feature extraction will be performed in the cloud based on available local resources; the indication information includes prompt words and necessary data.

[0052] In one embodiment, semantic features are extracted based on a public semantic knowledge base according to the instruction information, including:

[0053] In response to receiving prompts and necessary data, semantic features are extracted based on a public semantic knowledge base;

[0054] The semantic features are transmitted to the sending end.

[0055] In one embodiment, if the sending end determines that semantic feature extraction should be performed locally based on available local resources, it extracts semantic features based on a local semantic knowledge base.

[0056] In one embodiment, the method further includes:

[0057] Send optimization suggestions to the sender; the optimization suggestions are used to instruct the sender to optimize the semantic features extracted based on the local semantic knowledge base.

[0058] In one embodiment, before sending optimization suggestions to the sender, the method includes:

[0059] Receive prompts and necessary data from the sender;

[0060] Based on prompts and necessary data, optimization suggestions for semantic features are generated using a public semantic knowledge base.

[0061] In one embodiment, the prompt word is created by the sender based on semantic communication goals and semantic features.

[0062] In one embodiment, the method further includes:

[0063] The performance metrics for semantic feature extraction performed on the cloud are transmitted to the sending end. The performance metrics for semantic feature extraction performed on the cloud are used to instruct the sending end, based on the available local resources and the resources required for semantic feature extraction, to compare the performance metrics for semantic feature extraction performed locally with those for semantic feature extraction performed on the cloud, and to determine whether to perform semantic feature extraction locally or on the cloud based on the comparison results.

[0064] In one embodiment, the performance metrics for offloading semantic feature extraction to the cloud include cloud completion time, data transmission time for transferring the data to be processed to the server, and data transmission energy consumption; the method further includes:

[0065] Receive data volume information from the sender regarding the amount of data to be processed;

[0066] Based on the data volume information, calculate the cloud completion time, data transmission time, and data transmission energy consumption, and then transmit these data to the sending end.

[0067] In one embodiment, the data volume information includes at least one of data size and data type.

[0068] Thirdly, this application also provides a semantic feature extraction device, applied at the sending end, the device comprising:

[0069] The extraction method determination module is used to determine whether to perform semantic feature extraction locally or in the cloud, based on available local resources.

[0070] The local extraction module is used to extract semantic features based on the local semantic knowledge base if it is determined that semantic feature extraction will be performed locally.

[0071] The instruction module is used to send instruction information to the server if it is determined that semantic feature extraction will be performed in the cloud; the instruction information is used to instruct the server to extract semantic features based on a public semantic knowledge base.

[0072] Fourthly, this application also provides a semantic feature extraction apparatus, applied to a server, the apparatus comprising:

[0073] The information receiving module is used to receive indication information from the sending end;

[0074] The feature extraction module is used to extract semantic features based on a public semantic knowledge base according to the instruction information.

[0075] Fifthly, this application also provides a transmitting end, including: a transmitter and a processor;

[0076] The processor is used to determine whether to extract semantic features locally or in the cloud based on available local resources; if it is determined to extract semantic features locally, it extracts semantic features based on the local semantic knowledge base; and if it is determined to extract semantic features in the cloud, it controls the transmitter to send instruction information to the server; the instruction information is used to instruct the server to extract semantic features based on the public semantic knowledge base.

[0077] Sixthly, this application also provides a server, including: a receiver and a processor;

[0078] A receiver is used to receive indication information from the sender.

[0079] The processor is used to extract semantic features based on a public semantic knowledge base according to the instruction information.

[0080] In a seventh aspect, this application also provides a communication system, including the transmitting end described in the fifth aspect above, and the server described in the sixth aspect.

[0081] Eighthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the methods described above.

[0082] Ninthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the methods described above.

[0083] The aforementioned semantic feature extraction method, apparatus, device, and communication system allow the sending end to determine whether to perform semantic feature extraction locally or in the cloud based on available local resources. If local extraction is chosen, semantic features are extracted based on a local semantic knowledge base. If cloud extraction is chosen, an instruction is sent to the server, instructing the server to extract semantic features based on a public semantic knowledge base. This application determines the semantic feature extraction method based on available local resources, enabling end-to-cloud collaboration to assist the sending end in completing semantic feature extraction. This allows the semantic communication system to transmit information more effectively, thereby improving the quality and efficiency of the entire semantic communication process. Attached Figure Description

[0084] Figure 1 shows the application environment of a semantic feature extraction method in one embodiment;

[0085] Figure 2 is a flowchart illustrating a semantic feature extraction method in one embodiment;

[0086] Figure 3 is a schematic diagram of the process of determining whether to perform semantic feature extraction locally or in the cloud in one embodiment;

[0087] Figure 4 is a flowchart illustrating the process of determining the local capability to complete semantic feature extraction in one embodiment;

[0088] Figure 5 is a flowchart illustrating the process of determining that the available local resources are greater than the resources required for semantic feature extraction in one embodiment.

[0089] Figure 6 is a schematic diagram of the specific process for determining whether to perform semantic feature extraction locally or in the cloud in one embodiment;

[0090] Figure 7 is a flowchart illustrating the optimization of semantic features in one embodiment;

[0091] Figure 8 is a schematic diagram of the specific process for optimizing semantic features in one embodiment;

[0092] Figure 9 is a schematic diagram of the semantic feature extraction process in the cloud in one embodiment;

[0093] Figure 10 is a flowchart illustrating the semantic feature extraction method in another embodiment;

[0094] Figure 11 is a schematic diagram of a semantic feature extraction framework in one embodiment;

[0095] Figure 12 is a schematic diagram of the specific process of semantic feature extraction method in one embodiment;

[0096] Figure 13 is a structural block diagram of a semantic feature extraction device in one embodiment;

[0097] Figure 14 is a structural block diagram of a semantic feature extraction device in another embodiment;

[0098] Figure 15 is an internal structure diagram of a terminal in one embodiment;

[0099] Figure 16 is an internal structure diagram of the server in one embodiment. Detailed Implementation

[0100] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0101] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe the order or sequence of the objects. It should be understood that such data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, not limited in number; for example, the first object can be one or more. Furthermore, the term "and / or" merely describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0102] In the description of embodiments of this application, unless otherwise stated, "a plurality of" means two or more. When used herein, the singular forms "a," "an," and "the" may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms "comprising / including" or "having," etc., specify the presence of the stated features, integrals, steps, operations, components, parts, or combinations thereof, but do not preclude the possibility of the presence or addition of one or more other features, integrals, steps, operations, components, parts, or combinations thereof.

[0103] It is worth noting that the terms "system" and "network" in the embodiments of this application are often used interchangeably, and the described technologies can be used in the systems mentioned above as well as other systems. The following description describes a New Radio (NR) system for illustrative purposes, and the term NR is used in most of the following description. However, these technologies can also be applied to applications other than NR systems, such as 6th Generation (6G) communication systems, or to next-generation mobile communication systems or other similar communication systems, without limitation.

[0104] Figure 1 is a schematic diagram of a wireless communication system applicable to embodiments of this application. The wireless communication system includes at least one terminal 102 and a server 104. The terminal 102 communicates with the server 104 through a network. The terminal 102 and the server 104 can be connected to the network wirelessly or via a wired connection to complete data transmission and exchange.

[0105] Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can include virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc.

[0106] Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0107] For example, taking a cloud server (hereinafter referred to as a cloud server) as the server, the terminal, as the sending end, can be directly or indirectly connected to the cloud server through wired or wireless communication. This application embodiment does not impose any limitations.

[0108] In related technologies, feature extraction is performed using only edge-side capabilities. However, edge-side resources are usually limited. When faced with data-intensive or computationally intensive tasks, this may lead to significant processing latency or even the failure of the entire semantic communication process.

[0109] Based on the aforementioned technologies, this application embodiment addresses semantic communication systems by performing semantic feature extraction based on a semantic knowledge base. It leverages a cloud-based multimodal model to optimize edge-side features. Specifically, this application not only utilizes collaborative computing power between the edge and cloud to complete semantic feature extraction but also leverages the collaboration between the sender's local semantic knowledge base and the cloud-based public semantic knowledge base to optimize semantic features, thereby improving their accuracy. Furthermore, this application utilizes cloud-side and edge-side collaboration to complete semantic feature extraction, overcoming edge-side resource limitations and ensuring efficient semantic feature extraction even in resource-intensive tasks.

[0110] The semantic feature extraction method provided in this application ensures successful extraction of semantic features when faced with resource-intensive tasks through collaboration between the sending end and the cloud server. Simultaneously, it improves the accuracy of semantic feature extraction by leveraging the collaboration between the sending end's local semantic knowledge base and the cloud server's public semantic knowledge base.

[0111] It should be noted that the beneficial effects or technical problems solved by the embodiments of this application are not limited to this one, but may also be other implicit or related problems. For details, please refer to the description of the embodiments below.

[0112] Before introducing the specific embodiments of this application, the technical terms involved in this application will be explained:

[0113] Device-Cloud Collaboration: An artificial intelligence technology paradigm that allows cloud and edge devices to work together. It does not rely entirely on centralized cloud computing resources, but instead deploys some AI (Artificial Intelligence) tasks on the edge to solve the problems of privacy, security, load, and cost in traditional cloud intelligent services, and improve the real-time performance and personalized service capabilities of intelligent systems.

[0114] Semantic communication focuses on the meaning of transmitted information, rather than the precise reproduction of the communication symbols themselves. It is mainly used to solve the problem of how the sent symbols convey the exact meaning.

[0115] A semantic knowledge base is a structured knowledge network model that provides relevant semantic knowledge descriptions for data information and has memory capabilities. Semantic knowledge bases provide knowledge background and storage search services for the extraction, identification, transmission, understanding, and reasoning processes of semantic elements in semantic communication. They define an efficient search space, standardize search paths, and are one of the key enabling technologies for semantic communication. Typically, in a semantic communication system, the sending and receiving ends have their own local semantic knowledge bases, while a public semantic knowledge base is maintained on a cloud server.

[0116] Semantic features are key information used for semantic communication after selective feature extraction from raw information. They include not only the direct content of the data but also its context, relevance, and relationships with other data. The accuracy of semantic feature extraction directly affects whether the communication system can correctly understand and convey the intent and content of the information.

[0117] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with reference to one embodiment. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0118] In an exemplary embodiment, as shown in FIG2, a semantic feature extraction method is provided. Taking the application of this method to the terminal in FIG1 as an example (it is understood that the terminal can act as a sender to communicate with the server), the method includes the following steps 202 to 206. Wherein:

[0119] Step 202: Determine whether to perform semantic feature extraction locally or in the cloud, based on available local resources.

[0120] The locally available resources can be the resources currently available to the sending end (e.g., the terminal), such as computing resources, memory space, etc., which are not limited in this application.

[0121] In one embodiment, the sending end can determine the semantic feature extraction method based on the available local resources. The semantic feature extraction method in this embodiment may include a first extraction method and a second extraction method, wherein the second extraction method is different from the first extraction method. For example, the first extraction method may be to perform semantic feature extraction locally, while the second extraction method may be to perform semantic feature extraction in the cloud.

[0122] In one embodiment, the sending end can choose to perform semantic feature extraction in the cloud if the available local resources do not meet the corresponding conditions; or it can choose to perform semantic feature extraction locally if the available local resources meet the corresponding conditions, thereby enabling the sending end and the cloud to collaboratively complete the semantic feature extraction.

[0123] This application embodiment determines the semantic feature extraction method by utilizing locally available resources, enabling the cloud side and the edge side to collaboratively complete semantic feature extraction. This overcomes the limitations of edge side resources, allowing for the smooth extraction of semantic features when facing resource-intensive tasks, and thus ensuring efficient semantic feature extraction even in resource-intensive tasks.

[0124] Step 204: If it is determined that semantic feature extraction will be performed locally, then semantic features will be extracted based on the local semantic knowledge base.

[0125] In one embodiment, when it is determined that semantic feature extraction will be performed locally, the sending end can extract semantic features based on a local semantic knowledge base. The local semantic knowledge base can be the sending end's own semantic knowledge base.

[0126] For example, the sending end can extract semantic features based on a local semantic knowledge base if it determines that semantic feature extraction should be performed locally based on available local resources.

[0127] Step 206: If it is determined that semantic feature extraction will be performed in the cloud, then an instruction message is sent to the server; the instruction message is used to instruct the server to extract semantic features based on a public semantic knowledge base.

[0128] In one embodiment, when it is determined that semantic feature extraction will be performed in the cloud, the sending end can send an instruction message to the server, instructing the server to extract semantic features based on a public semantic knowledge base. The public semantic knowledge base can be a semantic knowledge base maintained by the server, such as a public semantic knowledge base maintained on a cloud server.

[0129] For example, taking a cloud server as an example, the sending end can send relevant data (such as instruction information) to the cloud server if it determines that semantic feature extraction will be performed in the cloud based on available local resources. The cloud server extracts semantic features based on a public semantic knowledge base.

[0130] The aforementioned semantic feature extraction method allows the sending end to determine the extraction method based on available local resources. This method can be applied to semantic communication systems, utilizing end-to-end cloud collaboration to assist the sending end in completing semantic feature extraction. This embodiment leverages the collaboration between the sending end's local semantic knowledge base and a public semantic knowledge base on a server (e.g., a cloud server) to extract semantic features, enabling the semantic communication system to transmit information more effectively, thereby improving the quality and efficiency of the entire semantic communication process.

[0131] In one embodiment, as shown in FIG3, step 202 may include steps 302 to 304. Wherein:

[0132] Step 302: Based on the available local resources and the resources required for semantic feature extraction, confirm whether the local system has the capability to complete semantic feature extraction.

[0133] In one embodiment, the sending end can determine whether it has the capability to complete semantic feature extraction based on locally available resources and the resources required for semantic feature extraction. The resources required for semantic feature extraction can refer to the resource requirements for semantic feature extraction. For example, the sending end can assess locally available resources and the resource requirements for semantic feature extraction to determine whether it has the capability to complete semantic feature extraction.

[0134] In one embodiment, the sending end can compare its local available resources with the resources required for semantic feature extraction to determine whether its local available resources can meet the resource requirements for semantic feature extraction, thereby confirming whether it has the capability to complete semantic feature extraction. In this embodiment, the sending end can determine the semantic feature extraction method based on its local available resources, thereby enabling the use of end-to-cloud collaboration to assist the sending end in completing semantic feature extraction.

[0135] In one embodiment, the sending end can determine available local resources through local resource assessment and obtain the resources required for semantic feature extraction through semantic feature extraction resource assessment. Furthermore, the sending end can also perform channel resource assessment to confirm the communication status between the sending end and the server. For example, in a semantic communication system, available channel resources are analyzed and assessed to ensure effective signal transmission and optimal network performance. Exemplarily, if the channel resource assessment determines that signal transmission is currently unsuitable, semantic feature extraction can be performed locally.

[0136] Step 304: If the local machine has the capability to complete semantic feature extraction, compare the performance metrics of completing semantic feature extraction locally with the performance metrics of completing semantic feature extraction on the cloud, and determine whether to perform semantic feature extraction locally or on the cloud based on the comparison results.

[0137] In one embodiment, if the local machine has the capability to perform semantic feature extraction, the sending end can compare the performance metrics of performing semantic feature extraction locally versus offloading it to the cloud, and then decide whether to perform semantic feature extraction locally or in the cloud based on the comparison results.

[0138] In this embodiment, the sending end can determine whether it has the capability to complete semantic feature extraction locally. If it does, it can then use an end-to-cloud collaborative approach to assist the sending end in completing semantic feature extraction, ensuring the smooth extraction of semantic features when facing resource-intensive tasks. Furthermore, it can utilize the collaboration between the sending end's local semantic knowledge base and the cloud server's public semantic knowledge base to optimize semantic features, thereby improving the accuracy of semantic feature extraction. This allows the semantic communication system to transmit information more effectively, thus improving the quality and efficiency of the entire semantic communication process.

[0139] In an exemplary embodiment, as shown in FIG4, step 302 may include steps 402 to 406. Wherein:

[0140] Step 402: Evaluate the data to be processed and obtain the data volume information of the data to be processed.

[0141] In one embodiment, the data to be processed can refer to the data that needs to be processed, wherein the sending end can evaluate the data to be processed and obtain the data volume information of the data to be processed.

[0142] For example, data volume information may include at least one of data size and data type. In one embodiment, the sending end may obtain the data size and data type of the data to be processed.

[0143] Step 404: Based on the data volume information, assess the resource requirements for running the local semantic feature extraction algorithm, and use the resource requirements as the resources needed for semantic feature extraction.

[0144] In one embodiment, the sending end can check whether there is a semantic feature extraction algorithm locally, and if it is determined that there is a semantic feature extraction algorithm locally, it can evaluate the resource requirements for running the algorithm. For example, the resource requirements may include at least one of computing requirements and memory requirements.

[0145] If the sending end assesses the resource requirements for running the local semantic feature extraction algorithm based on the data volume information, it can use those resource requirements as the resources needed for semantic feature extraction.

[0146] Step 406: If the available local resources are greater than the resources required for semantic feature extraction, then it is determined that the local system has the capability to complete semantic feature extraction.

[0147] In one embodiment, the sending end can determine whether the locally available resources are greater than the resources required to extract semantic features, for example, by comparing the locally available resources with the resources required to extract semantic features, to determine whether the locally available resources are greater than the resources required to extract semantic features.

[0148] If the available local resources are greater than the resources required for semantic feature extraction, the sending end can determine that it has the capability to complete semantic feature extraction; otherwise, it can determine that it does not have the capability to complete semantic feature extraction.

[0149] In one embodiment, taking resource requirements including computational and memory requirements as an example, in one embodiment, the resources required for semantic feature extraction may include computational requirements for semantic feature extraction and memory requirements for semantic feature extraction; as shown in Figure 5, step 406 may include steps 502 to 504. Wherein:

[0150] Step 502: Obtain locally available resources; locally available resources include locally available computing resources and locally available memory space.

[0151] In one embodiment, the sending end can obtain locally available resources, such as by checking locally available resources; wherein, locally available resources may include locally available computing resources and locally available memory space.

[0152] Step 504: If the available local computing resources are greater than the computational requirements for semantic feature extraction, and the available local memory space is greater than the memory requirements for semantic feature extraction, then it is determined that the local system has the capability to complete semantic feature extraction.

[0153] In one embodiment, the sending end determines whether the locally available resources are greater than the resources required to extract semantic features. If the locally available computing resources are greater than the computing requirements for semantic feature extraction, and the locally available memory space is greater than the memory requirements for semantic feature extraction, then the sending end has the capability to complete semantic feature extraction. Otherwise, it does not.

[0154] In practical applications, if the local machine has the capability to perform semantic feature extraction, the sending end can compare the performance metrics of performing semantic feature extraction locally and offloading it to the cloud, and decide whether to perform semantic feature extraction locally or in the cloud based on the comparison results. In an exemplary embodiment, as shown in Figure 6, step 504 may include steps 602 to 606. Wherein:

[0155] Step 602: Transmit the data volume information to the server; the data volume information is used to instruct the server to calculate and provide feedback on the cloud completion time, data transmission time, and data transmission energy consumption.

[0156] In one embodiment, the performance metrics for locally performed semantic feature extraction may include local completion time and local completion energy consumption, where local completion time represents the time required to perform semantic feature extraction locally, and local completion energy consumption represents the energy required to perform semantic feature extraction locally. For example, the sending end may calculate the time and energy required to complete semantic feature extraction.

[0157] In one embodiment, the performance metrics for offloading semantic feature extraction to the cloud may include cloud completion time, data transmission time for transmitting the data to be processed to the server, and data transmission energy consumption; wherein, cloud completion time represents the time required to complete semantic feature extraction in the cloud.

[0158] In this embodiment, the sending end can transmit data volume information to the server to instruct the server to calculate the cloud completion time, data transmission time, and data transmission energy consumption, and then feed back these data to the sending end. Taking a cloud server as an example, the sending end can send the data size and data type to the cloud server. The cloud server calculates the time and energy consumption required to complete semantic feature extraction and transmits the results to the sending end. The sending end compares the time and energy consumption for local completion with those for cloud completion, and decides whether to perform semantic feature extraction locally or in the cloud.

[0159] Step 604: If the local completion time is less than the sum of the data transmission time and the cloud completion time, and the local completion energy consumption is less than the data transmission energy consumption, then determine to perform semantic feature extraction locally.

[0160] In one embodiment, if the local completion time is less than the data transmission time plus the cloud completion time, and the local completion energy consumption is less than the data transmission energy consumption, then it is determined that semantic feature extraction will be performed locally.

[0161] Step 606: If the local completion time is greater than or equal to the sum of the data transmission time and the cloud completion time, and / or the local completion energy consumption is greater than or equal to the data transmission energy consumption, then determine to perform semantic feature extraction in the cloud.

[0162] In one embodiment, if the local completion time is greater than or equal to the sum of the data transmission time and the cloud completion time, and / or the local completion energy consumption is greater than or equal to the data transmission energy consumption, then the sending end determines to perform semantic feature extraction in the cloud.

[0163] It is understandable that, taking a terminal as an example, the sending end typically relies on battery power, and its battery capacity is limited. When the terminal completes complex tasks, special attention needs to be paid to the energy consumption of the terminal device. In contrast, the cloud has higher energy efficiency and larger energy capacity. Based on this, the embodiments of this application determine whether to perform semantic feature extraction locally or in the cloud by comparing the performance indicators of semantic feature extraction performed locally and offloading it to the cloud, thereby realizing the use of end-cloud collaboration to assist the sending end in completing semantic feature extraction.

[0164] The aforementioned semantic feature extraction method ensures successful extraction of semantic features even under resource-intensive tasks through collaboration between the sending end and the cloud server. Simultaneously, it improves the accuracy of semantic feature extraction by leveraging the collaboration between the sending end's local semantic knowledge base and the cloud server's public semantic knowledge base.

[0165] In an exemplary embodiment, as shown in FIG7, step 204 may include steps 702 to 704. Wherein:

[0166] Step 702: Extract semantic features based on the local semantic knowledge base.

[0167] In one embodiment, the semantic features extracted by the sending end can refer to the semantic features extracted by the sending end based on a local semantic knowledge base. Specifically, the sending end can extract semantic features based on a local semantic knowledge base, provided it is determined that semantic feature extraction will be performed locally. It is understood that, for ease of distinction, the semantic features extracted by the sending end based on a local semantic knowledge base in this embodiment can be referred to as the first semantic features.

[0168] Step 704: Receive optimization suggestions from the server and optimize the semantic features according to the optimization suggestions.

[0169] In one embodiment, the sending end can receive optimization suggestions transmitted by the server, and then optimize the semantic features according to the optimization suggestions. For example, the optimization suggestions can be used to instruct the sending end to optimize the semantic feature results (i.e., semantic features) extracted locally, so as to achieve more accurate semantic feature extraction.

[0170] In one embodiment, the optimization suggestion can be obtained by the server based on a public semantic knowledge base. For example, the sending end can send semantic features and related data extracted from a local semantic knowledge base to the server, so that the server can obtain optimization suggestions based on the public semantic knowledge base. It is understood that the embodiments of this application do not limit the way the server obtains optimization suggestions.

[0171] In an exemplary embodiment, as shown in FIG8, steps 802 to 804 may be included before step 704. Wherein:

[0172] Step 802: Create prompt words based on semantic communication goals and semantic features.

[0173] In one embodiment, once semantic features are acquired, the sending end can create prompt words based on the semantic communication target and the semantic features. The semantic communication target can refer to the target information that needs to be obtained through this semantic communication in the current application scenario; taking an autonomous driving scenario as an example, in an autonomous driving scenario, the semantic communication target can be obtaining surrounding environmental information, including pedestrians, vehicles, traffic signals, etc., through semantic communication between vehicles and between vehicles and traffic infrastructure.

[0174] For example, depending on the specific application scenario (e.g., semantic communication scenario), the semantic communication target can be obtained by mutual confirmation between the two parties before the formal communication process, or it can be obtained from a server (e.g., a cloud server).

[0175] It is understood that, for ease of distinction, the prompt word created by the sending end based on semantic communication goals and semantic features in the embodiments of this application can be referred to as the first prompt word.

[0176] Step 804: The prompt words and necessary data are transmitted to the server to instruct the server to generate optimization suggestions for semantic features based on a public semantic knowledge base, and then transmit the optimization suggestions to the sender.

[0177] In one embodiment, the sending end can transmit prompt words and necessary data to the server, thereby instructing the server to generate optimization suggestions for semantic features based on a public semantic knowledge base, and then transmit the optimization suggestions to the sending end.

[0178] Here, necessary data may refer to data related to semantic feature extraction. In one embodiment, when it is determined that semantic feature extraction is performed locally, the necessary data may include semantic features extracted by the sending end based on the local semantic knowledge base. It can be understood that, for ease of distinction, the necessary data used to instruct the server to generate optimization suggestions in this application embodiment may be referred to as first necessary data.

[0179] In one embodiment, optimization suggestions can be used to instruct the sender to optimize the semantic feature results (i.e., semantic features) extracted locally. For example, taking a cloud server as an example, the sender can transmit prompt words and necessary data to the cloud server. The cloud server receives the prompt words and necessary data, generates optimization suggestions for the semantic features based on a public semantic knowledge base, and transmits the optimization suggestions to the sender.

[0180] In this embodiment, by transmitting prompt words and necessary data to the server, when semantic feature extraction is determined to be performed locally, the sending end can integrate semantic communication goals, create and transmit prompt words to guide the cloud to more accurately understand the semantic communication scenario and the sending end's intent, thereby generating more accurate optimization suggestions or achieving more accurate semantic feature extraction. This embodiment enables cloud processing and optimization to be more closely matched with the sending end's personalized needs and context, thereby improving the accuracy and efficiency of the entire semantic feature extraction process.

[0181] Taking a cloud server as an example, where the sending end determines to perform semantic feature extraction locally, the sending end extracts semantic features based on its local semantic knowledge base. Based on the semantic communication target and the extracted semantic features, the sending end creates prompt words. The sending end transmits the prompt words and necessary data to the cloud server. The cloud server receives the prompt words and necessary data, generates optimization suggestions for the semantic features based on a public semantic knowledge base, and transmits the optimization suggestions to the sending end. The sending end receives the suggestions from the cloud server and optimizes the semantic feature results. This embodiment of the application utilizes the collaboration between the sending end's local semantic knowledge base and the cloud server's public semantic knowledge base to improve the accuracy of semantic feature extraction.

[0182] The aforementioned semantic feature extraction method allows the sending end to determine the extraction method based on available local resources. It can also leverage both a local semantic knowledge base on the sending end and a public semantic knowledge base on the cloud side to extract and optimize semantic features, ultimately improving the accuracy of semantic feature extraction. Specifically, this application embodiment can be applied to semantic communication systems, using an end-cloud collaborative approach to assist the sending end in completing semantic feature extraction. In one embodiment, by utilizing the collaboration between the sending end's local semantic knowledge base and the cloud server's public semantic knowledge base, semantic feature extraction and optimization are completed, improving the accuracy of semantic feature extraction. This allows the semantic communication system to transmit information more effectively, thereby improving the quality and efficiency of the entire semantic communication process.

[0183] In one embodiment, the necessary data includes one or more of the following: semantic features, raw data, data type, data format, data creation time, and data version.

[0184] In one embodiment, when it is confirmed that semantic feature extraction is performed locally, the necessary data may include the semantic features extracted by the sending end; in another embodiment, the necessary data may further include at least one of the following: raw data, data type, data format, data creation time, and data version. The raw data may be data collected by the sending end through sensors and then locally anonymized on the device side; the data type may be the modality of the data collected by the sending end, such as text, voice, image, video, etc.

[0185] For example, taking a cloud server as the server, if it is determined that semantic feature extraction is performed locally, the sending end can extract semantic features based on the local semantic knowledge base, and then create prompt words based on the semantic communication target and the extracted semantic features; the sending end transmits the prompt words and necessary data to the cloud server, wherein the necessary data may include one or more of the following: semantic features, original data, data type, data format, data creation time, and data version.

[0186] In one embodiment, in an exemplary embodiment, the prompt word includes one or more of a description of the communication target, a summary of semantic features, and the sender's intent.

[0187] In one embodiment, when it is confirmed that semantic feature extraction is performed locally and the sending end extracts semantic features based on the local semantic knowledge base, the prompt word (i.e., the first prompt word) created by the sending end based on the semantic communication goal and the extracted semantic features may include one or more of the following: a description of the communication goal, a summary of the semantic features, and the sending end's intent.

[0188] For example, the communication target description can refer to a description of the semantic communication target; the semantic feature summary can be a semantic-level extraction and representation of semantic features, such as keywords, topics, etc.; the sender's intent can be an operation that needs to be performed by the cloud, information that needs to be understood, or a behavior that needs to be triggered.

[0189] In one embodiment, the cue words are presented in one or more of the following formats: text, vector, and structured data.

[0190] In one embodiment, the presentation of the prompt words includes, but is not limited to, text form, vector form, and structured data. For example, the prompt words can be presented in text form, or in vector form or structured data; this application is not limited in this regard.

[0191] In some embodiments, the local semantic knowledge base includes one or more of user information, device information, target domain knowledge, and contextual information.

[0192] In one embodiment, the local semantic knowledge base of the sending end may include at least one of user information, device information, target domain knowledge, and context information. User information may refer to information related to user identity and usage patterns; device information may refer to information related to device attributes, device configuration, and operation; target domain knowledge may refer to structured knowledge describing information, concepts, and their interrelationships within a specific target domain; and context information may be information related to the application environment of the sending end.

[0193] In one embodiment, user information may include one or more of user profiles, user preferences, user behavior patterns, and historical interaction records; device information may include one or more of hardware configuration, software configuration, performance parameters, and device operating status; target domain knowledge may include one or more of concept definitions, entity relationships, rule logic, and domain models; and context information may include one or more of environmental state, situational information, and time information.

[0194] In one embodiment, semantic features include one or more of entity recognition, sentiment analysis, topic classification, and semantic role labeling.

[0195] In one embodiment, when it is confirmed that semantic feature extraction is performed locally, the semantic features extracted by the sending end based on the local semantic knowledge base (i.e., semantic features) may include one or more of entity recognition, sentiment analysis, topic classification, and semantic role labeling. Entity recognition is used to represent entities that represent the target meaning, such as names of people, places, organizations, dates, and proper nouns; sentiment analysis can refer to identifying the emotional tendency expressed by the input data (such as positive, negative, or neutral, or more specifically, positivity, negativity, and neutrality); topic classification can refer to categorizing the input data into predefined topics or categories; and semantic role labeling can refer to assigning semantic roles to various components in the input data, such as the executor of an action, the receiver of an action, etc.

[0196] Regarding the public semantic knowledge base and optimization suggestions, in some embodiments, the public semantic knowledge base may include one or more of general knowledge, domain expertise, device interaction knowledge, and model libraries; optimization suggestions may include one or more of maintaining, expanding, deleting, merging, splitting, refining, and simplifying.

[0197] In one embodiment, the public semantic knowledge base maintained by the server may include at least one of general knowledge, domain-specific knowledge, device interaction knowledge, and a model library; wherein, general knowledge may refer to general knowledge, domain-specific knowledge may refer to professional knowledge of the target domain, and device interaction knowledge may involve interaction information between devices, including device functions, interface standards, communication protocols, etc. The model library may contain various models and algorithms.

[0198] In one embodiment, general knowledge includes one or more of common sense, conceptual definitions, and general rules; domain expertise includes one or more of technical terms, cases, regulations, and industry standards; device interaction knowledge includes one or more of device types, interfaces, communication protocols, and operating instructions; and the model library includes one or more of statistical models, machine learning models, and deep learning models.

[0199] In one embodiment, the optimization suggestions in this application embodiment may include at least one of maintaining, expanding, deleting, merging, splitting, refining, and simplifying. After receiving the optimization suggestions transmitted by the server, the sending end can optimize the semantic feature results. For example, it can perform operations such as maintaining, expanding, deleting, merging, splitting, refining, and / or simplifying on the semantic features extracted based on the local semantic knowledge base to improve the accuracy of semantic feature extraction.

[0200] It is understood that when semantic feature extraction is determined to be performed in the cloud, the sending end can send relevant data (e.g., instruction information) to the server to instruct the server to extract semantic features based on a public semantic knowledge base. In one embodiment, the instruction information may include prompt words and necessary data.

[0201] The prompts and necessary data transmitted from the sending end to the server can be used to instruct the server to extract semantic features based on a public semantic knowledge base.

[0202] In some embodiments, as shown in FIG9, step 206 may include steps 902 to 904. Wherein:

[0203] Step 902: Create prompt words based on semantic communication goals and the local semantic knowledge base.

[0204] In one embodiment, when it is determined that semantic feature extraction will be performed in the cloud, the sending end can create prompt words based on the semantic communication target and the local semantic knowledge base. It is understood that, for ease of distinction, the prompt words created by the sending end based on the semantic communication target and the local semantic knowledge base in this embodiment can be referred to as second prompt words.

[0205] The semantic communication target refers to the target information that needs to be obtained through this semantic communication in the current application scenario. Taking the autonomous driving scenario as an example, the semantic communication target can be to obtain information about the surrounding environment, including pedestrians, vehicles, traffic signals, etc., through semantic communication between vehicles and between vehicles and traffic infrastructure.

[0206] In one embodiment, depending on the specific application scenario, the semantic communication target may be obtained by mutual confirmation between the two parties before the formal communication process, or it may be obtained from a server (e.g., a cloud server).

[0207] Step 904: The prompt words and necessary data are transmitted to the server to instruct the server to extract semantic features based on the public semantic knowledge base and transmit the semantic features to the sending end.

[0208] In one embodiment, when it is determined that semantic feature extraction will be performed in the cloud, the sending end can send instruction information to the server, wherein the instruction information may include prompt words and necessary data. Then, upon receiving the prompt words and necessary data, the server can extract semantic features based on a public semantic knowledge base and transmit the semantic features to the sending end.

[0209] It is understood that, for ease of distinction, the necessary data used to instruct the server to extract semantic features based on the public semantic knowledge base in this application embodiment can be referred to as the second necessary data. The semantic features extracted by the server based on the public semantic knowledge base in this application embodiment can be referred to as the second semantic features.

[0210] Taking a cloud server as an example, when it is determined that semantic feature extraction will be performed in the cloud, the sending end creates prompt words based on the semantic communication target and the local semantic knowledge base; the sending end transmits the prompt words and necessary data to the cloud server; the cloud server receives the prompt words and necessary data, extracts semantic features based on the public semantic knowledge base; and the cloud server transmits the extracted semantic features back to the sending end.

[0211] In summary, once semantic features are determined to be extracted in the cloud, the sending end can integrate semantic communication goals, create and transmit prompt words to guide the cloud to more accurately understand the semantic communication scenario and the sending end's intent. This allows for the generation of more precise optimization suggestions or the achievement of more accurate semantic feature extraction. By sending instruction information to the cloud, the cloud's processing and optimization efforts can be more closely aligned with the sending end's personalized needs and context, thereby improving the accuracy and efficiency of the entire semantic feature extraction process.

[0212] In one embodiment, the necessary data includes at least one of the following: raw data, data type, data format, data creation time, and data version.

[0213] In one embodiment, when it is determined that semantic features will be extracted in the cloud, the necessary data may include one or more of the following: raw data, data type, data format, data creation time, and data version. The raw data may be data collected by the sending end through sensors and then locally anonymized on the device. The data type may be the modality of the data collected by the sending end, such as text, voice, image, or video.

[0214] In some embodiments, the prompt word includes one or more of the following: a description of the communication target, a description of the communication scenario, the sender's intent, and a personalized tag.

[0215] In one embodiment, when it is determined that semantic features are extracted in the cloud, the prompt words created by the sending end based on the semantic communication target and the local semantic knowledge base may include at least one of the following: communication target description, communication scenario description, sending end intent, and personalized tags. The communication target description may be a description of the semantic communication target, the communication scenario description may be a description of the current application scenario, the sending end intent may be an operation that needs to be performed in the cloud, information that needs to be understood, or a behavior that needs to be triggered, and the personalized tags may be identifiers generated based on relevant information of the semantic communication target, used to describe target features or content attributes.

[0216] In some embodiments, the cue words are presented in one or more of the following formats: text, vector, and structured data.

[0217] In one embodiment, the presentation of the prompt words includes, but is not limited to, text form, vector form, and structured data. For example, the prompt words can be presented in text form, or in vector form or structured data; this application is not limited in this regard.

[0218] It should be noted that when semantic feature extraction is performed in the cloud, the forms used by the aforementioned local semantic database and public semantic database can be found in the limitations on local semantic databases and public semantic databases in the section on semantic feature extraction in the local context, and will not be repeated here.

[0219] The aforementioned semantic feature extraction method, whether extracting semantic features locally or in the cloud, allows the sending end to integrate semantic communication goals, create and transmit prompt words, and guide the cloud to more accurately understand the semantic communication scenario and the sending end's intent. This leads to the generation of more precise optimization suggestions or more accurate semantic feature extraction. The embodiments of this application enable cloud processing and optimization to be more closely matched with the sending end's personalized needs and context, thereby improving the accuracy and efficiency of the entire semantic feature extraction process.

[0220] In an exemplary embodiment, as shown in FIG10, a semantic feature extraction method is provided. Taking the application of this method to the server in FIG1 as an example (it is understood that the server can be a cloud server), the method includes the following steps 1002 to 1004. Wherein:

[0221] Step 1002: Receive indication information from the sending end.

[0222] In one embodiment, the server may perform semantic feature extraction in response to receiving an indication from the sender. This indication is sent by the sender if it determines, based on available local resources, that semantic feature extraction will be performed in the cloud.

[0223] Step 1004: Extract semantic features based on the public semantic knowledge base according to the instruction information.

[0224] In one embodiment, the server can extract semantic features based on a public semantic knowledge base according to the instruction information.

[0225] In one embodiment, regarding the execution of semantic feature extraction in the cloud, in some embodiments, the indication information is sent by the sending end when it determines that semantic feature extraction will be performed in the cloud based on locally available resources; wherein, the indication information includes prompt words and necessary data. In the embodiments of this application, the server can extract semantic features based on a public semantic knowledge base according to the prompt words and necessary data sent by the sending end.

[0226] In one embodiment, semantic features are extracted based on a public semantic knowledge base according to the instruction information, including:

[0227] In response to receiving prompts and necessary data, semantic features are extracted based on a public semantic knowledge base;

[0228] The semantic features are transmitted to the sending end.

[0229] In one embodiment, the server receives the prompt words and necessary data, extracts semantic features based on a public semantic knowledge base, and transmits the semantic features to the sender.

[0230] In some embodiments, the necessary data includes at least one of the following: raw data, data type, data format, data creation time, and data version.

[0231] In one embodiment, regarding the semantic feature extraction performed in the cloud, the necessary data sent by the sending end to the server may include one or more of the following: raw data, data type, data format, data creation time, and data version.

[0232] In one exemplary embodiment, the prompt word includes one or more of the following: a description of the communication goal, a description of the communication scenario, the sender's intent, and a personalized tag.

[0233] In one embodiment, the cue words are presented in one or more of the following formats: text, vector, and structured data.

[0234] In one exemplary embodiment, if the sending end determines that semantic feature extraction should be performed locally based on available local resources, it extracts semantic features based on a local semantic knowledge base.

[0235] In one embodiment, the method further includes:

[0236] Send optimization suggestions to the sender; the optimization suggestions are used to instruct the sender to optimize the semantic features extracted based on the local semantic knowledge base.

[0237] In one embodiment, when the sending end determines to perform semantic feature extraction locally, the server can send optimization suggestions to the sending end, and then the sending end optimizes the semantic features extracted based on the local semantic knowledge base.

[0238] In one embodiment, optimization suggestions may include one or more of maintaining, expanding, reducing, merging, splitting, refining, and simplifying.

[0239] In one embodiment, before sending optimization suggestions to the sender, the following is included:

[0240] Receive prompts and necessary data from the sender;

[0241] Based on prompts and necessary data, optimization suggestions for semantic features are generated using a public semantic knowledge base.

[0242] In some embodiments, the prompt word is created by the sender based on semantic communication goals and semantic features.

[0243] In one embodiment, when the sender determines to perform semantic feature extraction locally, the prompt word sent by the sender to the server can be created by the sender based on the semantic communication goal and semantic features.

[0244] In one embodiment, the cue word includes one or more of a description of the communication goal, a summary of semantic features, and the sender's intent.

[0245] In some embodiments, the cue words are presented in one or more of the following formats: text, vector, and structured data.

[0246] In one embodiment, the necessary data includes one or more of the following: semantic features, raw data, data type, data format, data creation time, and data version.

[0247] In one embodiment, when the sending end determines to perform semantic feature extraction locally, the necessary data sent by the sending end to the server may include one or more of the following: semantic features, raw data, data type, data format, data creation time, and data version. The semantic features included in the necessary data are semantic features extracted by the sending end based on the local semantic knowledge base.

[0248] In one embodiment, the local semantic knowledge base includes one or more of user information, device information, target domain knowledge, and contextual information.

[0249] In one embodiment, user information includes one or more of user profiles, user preferences, user behavior patterns, and historical interaction records; device information includes one or more of hardware configuration, software configuration, performance parameters, and device operating status; target domain knowledge includes one or more of concept definitions, entity relationships, rule logic, and domain models; and context information includes one or more of environmental state, situational information, and time information.

[0250] In one embodiment, semantic features include one or more of entity recognition, sentiment analysis, topic classification, and semantic role labeling.

[0251] In one embodiment, the public semantic knowledge base includes one or more of general knowledge, domain expertise, device interaction knowledge, and a model library.

[0252] In one embodiment, general knowledge includes one or more of common sense, conceptual definitions, and general rules; domain expertise includes one or more of technical terms, cases, regulations, and industry standards; device interaction knowledge includes one or more of device types, interfaces, communication protocols, and operating instructions; and the model library includes one or more of statistical models, machine learning models, and deep learning models.

[0253] In one embodiment, the method further includes:

[0254] The performance metrics for semantic feature extraction performed on the cloud are transmitted to the sending end. The performance metrics for semantic feature extraction performed on the cloud are used to instruct the sending end, based on the available local resources and the resources required for semantic feature extraction, to compare the performance metrics for semantic feature extraction performed locally with those for semantic feature extraction performed on the cloud, and to determine whether to perform semantic feature extraction locally or on the cloud based on the comparison results.

[0255] In one embodiment, the performance metrics for offloading semantic feature extraction to the cloud include cloud completion time, data transmission time for transferring the data to be processed to the server, and data transmission energy consumption; the method further includes:

[0256] Receive data volume information from the sender regarding the amount of data to be processed;

[0257] Based on the data volume information, calculate the cloud completion time, data transmission time, and data transmission energy consumption, and then transmit these data to the sending end.

[0258] In one embodiment, the data volume information includes at least one of data size and data type.

[0259] It is understood that the specific limitations of the semantic feature extraction method embodiment executed from the server can be found in the limitations of the semantic feature extraction method executed from the sending end above, and will not be repeated here.

[0260] In one embodiment, Figure 11 provides an interaction flowchart between a cloud server and a sending end (e.g., a terminal) in a semantic feature extraction method, as shown in Figure 12, which may include the following steps.

[0261] 1) The sending end assesses the resource requirements for semantic feature extraction and the available local resources. Based on the assessment results, the sending end determines whether it has the capability to complete semantic feature extraction locally; specifically, the sending end can determine its capability to complete semantic feature extraction locally by assessing the available local resources, channel resources, and the resource requirements for semantic feature extraction.

[0262] 2) If the local machine has the capability to perform semantic feature extraction, the sending end can compare the performance metrics of performing semantic feature extraction locally with those of performing it on the cloud, and decide whether to perform semantic feature extraction locally or on the cloud based on the comparison result. The comparison result can indicate whether the performance metrics of extracting semantic features locally are superior to those of extracting them on the cloud.

[0263] 3) If semantic feature extraction is performed locally at the sending end, then:

[0264] a) The sending end extracts semantic features based on a local semantic knowledge base;

[0265] b) The sending end creates prompt words based on the semantic communication goal and the extracted semantic features;

[0266] c) The sending end transmits the prompt and necessary data to the cloud server;

[0267] d) The cloud server receives prompt words and necessary data, and generates optimization suggestions for semantic features based on a public semantic knowledge base;

[0268] e) The sending end receives optimization suggestions from the cloud server and optimizes the semantic features.

[0269] 4) If semantic feature extraction is performed locally in the cloud, then:

[0270] a) The sending end creates prompt words based on the semantic communication target and the local semantic knowledge base;

[0271] b) The sending end transmits the prompt and necessary data to the cloud server;

[0272] c) The cloud server receives prompt words and necessary data, and extracts semantic features based on a public semantic knowledge base;

[0273] d) The cloud server transmits the extracted semantic features to the sending end.

[0274] The following example illustrates this concept using an autonomous driving scenario. In an autonomous driving scenario, vehicle A extracts semantic features based on a local semantic knowledge base. These extracted features include time, weather, road type, road conditions, vehicle status, and traffic lights. Based on the semantic communication objective and the extracted semantic features, vehicle A creates a prompt: "Currently on a certain branch of a certain highway, there are no vehicles ahead, and the reflector on the left side of the road is interrupted. Can we continue driving normally?" Vehicle A sends the prompt and a road photo to the cloud server. Upon receiving the prompt and the road photo, the cloud server, based on a public semantic knowledge base, generates an optimization suggestion for the semantic features: "The road on the left ahead has collapsed." The cloud server then sends this optimization suggestion to vehicle A.

[0275] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0276] Based on the same inventive concept, this application also provides a semantic feature extraction device for implementing the semantic feature extraction method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more semantic feature extraction device embodiments provided below can be found in the limitations of the semantic feature extraction method described above, and will not be repeated here.

[0277] In an exemplary embodiment, as shown in FIG13, a semantic feature extraction device is provided, applied at a sending end, the device comprising:

[0278] The extraction method determination module 1101 is used to determine whether semantic feature extraction should be performed locally or in the cloud based on available local resources.

[0279] The local extraction module 1102 is used to extract semantic features based on the local semantic knowledge base if it is determined that semantic feature extraction will be performed locally.

[0280] The instruction module 1103 is used to send instruction information to the server if it is determined that semantic feature extraction will be performed in the cloud; the instruction information is used to instruct the server to extract semantic features based on a public semantic knowledge base.

[0281] In one embodiment, the extraction method determination module 1101 includes:

[0282] The capability determination module is used to determine whether the local system has the capability to complete semantic feature extraction based on available local resources and the resources required for semantic feature extraction.

[0283] The metrics comparison module is used to compare the performance metrics of semantic feature extraction performed locally with those performed on the cloud if the local machine has the capability to perform semantic feature extraction. Based on the comparison results, it determines whether to perform semantic feature extraction locally or on the cloud.

[0284] In one embodiment, the capability determination module is used to assess the data to be processed and obtain the data volume information of the data to be processed; based on the data volume information, assess the resource requirements for running the local semantic feature extraction algorithm and use the resource requirements as the resources required for semantic feature extraction; if the available local resources are greater than the resources required for semantic feature extraction, then it is determined that the local system has the capability to complete semantic feature extraction.

[0285] In one embodiment, the data volume information includes at least one of data size and data type; the resource requirements include at least one of computing requirements and memory requirements.

[0286] In one embodiment, the resources required for semantic feature extraction include semantic feature extraction computational requirements and semantic feature extraction memory requirements;

[0287] The capability determination module is used to obtain local available resources; local available resources include local available computing resources and local available memory space; if the local available computing resources are greater than the semantic feature extraction computing requirements and the local available memory space is greater than the semantic feature extraction memory requirements, then it is determined that the local system has the capability to complete semantic feature extraction.

[0288] In one embodiment, the performance metrics for semantic feature extraction performed locally include local completion time and local completion energy consumption; the performance metrics for semantic feature extraction performed offloaded to the cloud include cloud completion time, data transmission time and data transmission energy consumption for transmitting the data to be processed to the server.

[0289] The indicator comparison module is used to transmit data volume information to the server. The data volume information is used to instruct the server to calculate and return the cloud completion time, data transmission time, and data transmission energy consumption. If the local completion time is less than the sum of the data transmission time and the cloud completion time, and the local completion energy consumption is less than the data transmission energy consumption, then it is determined to perform semantic feature extraction locally. If the local completion time is greater than or equal to the sum of the data transmission time and the cloud completion time, and / or the local completion energy consumption is greater than or equal to the data transmission energy consumption, then it is determined to perform semantic feature extraction in the cloud.

[0290] In one embodiment, the local extraction module 1102 includes:

[0291] The local feature extraction module is used to extract semantic features based on the local semantic knowledge base;

[0292] The optimization module receives optimization suggestions from the server and optimizes the semantic features based on these suggestions.

[0293] In one embodiment, the local extraction module 1102 further includes:

[0294] The prompt creation module is used to create prompt words based on semantic communication goals and semantic features;

[0295] The data transmission module is used to transmit prompt words and necessary data to the server, instructing the server to generate optimization suggestions based on semantic features according to a public semantic knowledge base, and then transmit the optimization suggestions to the sending end.

[0296] In one embodiment, the necessary data includes one or more of the following: semantic features, raw data, data type, data format, data creation time, and data version.

[0297] In one embodiment, the prompt word includes one or more of a communication goal description, a semantic feature summary, and the sender's intent.

[0298] In one embodiment, the prompts are presented in one or more of the following formats: text, vector, and structured data.

[0299] In one embodiment, the local semantic knowledge base includes one or more of user information, device information, target domain knowledge, and contextual information.

[0300] In one embodiment, user information includes one or more of user profiles, user preferences, user behavior patterns, and historical interaction records; device information includes one or more of hardware configuration, software configuration, performance parameters, and device operating status; target domain knowledge includes one or more of concept definitions, entity relationships, rule logic, and domain models; and context information includes one or more of environmental state, situational information, and time information.

[0301] In one embodiment, semantic features include one or more of entity recognition, sentiment analysis, topic classification, and semantic role labeling.

[0302] In one embodiment, the public semantic knowledge base includes one or more of general knowledge, domain expertise, device interaction knowledge, and a model library.

[0303] In one embodiment, general knowledge includes one or more of common sense, conceptual definitions, and general rules; domain expertise includes one or more of technical terms, cases, regulations, and industry standards; device interaction knowledge includes one or more of device types, interfaces, communication protocols, and operating instructions; and the model library includes one or more of statistical models, machine learning models, and deep learning models.

[0304] In one embodiment, optimization suggestions include one or more of maintaining, expanding, reducing, merging, splitting, refining, and simplifying.

[0305] In one embodiment, the instruction information includes prompts and necessary data.

[0306] In one embodiment, the indicating module 1003 includes:

[0307] The prompt word creation module is used to create prompt words based on semantic communication goals and a local semantic knowledge base;

[0308] The transmission module is used to transmit prompt words and necessary data to the server, instructing the server to extract semantic features based on a public semantic knowledge base and transmit the semantic features to the sending end.

[0309] In one embodiment, the necessary data includes at least one of the following: raw data, data type, data format, data creation time, and data version.

[0310] In one embodiment, the prompt word includes one or more of the following: a description of the communication goal, a description of the communication scenario, the sender's intent, and a personalized tag;

[0311] The cue words can be presented in one or more of the following formats: text, vector, and structured data.

[0312] In an exemplary embodiment, as shown in FIG14, a semantic feature extraction apparatus is provided, applied to a server, the apparatus comprising:

[0313] The information receiving module 1201 is used to receive indication information from the sending end;

[0314] The feature extraction module 1202 is used to extract semantic features based on a public semantic knowledge base according to the instruction information.

[0315] In one embodiment, the indication information is sent by the sending end when it determines that semantic feature extraction will be performed in the cloud based on available local resources; the indication information includes prompt words and necessary data.

[0316] In one embodiment, the feature extraction module 1102 is used to extract semantic features based on a public semantic knowledge base in response to receiving prompt words and necessary data; and to transmit the semantic features to the sending end.

[0317] In one embodiment, the necessary data includes at least one of the following: raw data, data type, data format, data creation time, and data version.

[0318] In one embodiment, the prompt word includes one or more of the following: a description of the communication goal, a description of the communication scenario, the sender's intent, and a personalized tag.

[0319] In one embodiment, the prompt words are presented in one or more of the following formats: text, vector, and structured data.

[0320] In one embodiment, if the sending end determines that semantic feature extraction should be performed locally based on available local resources, it extracts semantic features based on a local semantic knowledge base.

[0321] In one embodiment, the device further includes:

[0322] The optimization module is used to send optimization suggestions to the sending end; the optimization suggestions are used to instruct the sending end to optimize the semantic features extracted based on the local semantic knowledge base.

[0323] In one embodiment, the optimization module includes:

[0324] The data receiving module is used to receive prompts and necessary data from the sending end;

[0325] The optimization suggestion generation module is used to generate optimization suggestions for semantic features based on prompt words and necessary data, using a public semantic knowledge base.

[0326] In one embodiment, the prompt word is created by the sender based on semantic communication goals and semantic features.

[0327] In one embodiment, the necessary data includes one or more of the following: semantic features, raw data, data type, data format, data creation time, and data version.

[0328] In one embodiment, the prompt word includes one or more of a communication goal description, a semantic feature summary, and the sender's intent.

[0329] In one embodiment, the prompts are presented in one or more of the following formats: text, vector, and structured data.

[0330] In one embodiment, semantic features include one or more of entity recognition, sentiment analysis, topic classification, and semantic role labeling.

[0331] In one embodiment, optimization suggestions include one or more of maintaining, expanding, reducing, merging, splitting, refining, and simplifying.

[0332] In one embodiment, the local semantic knowledge base includes one or more of user information, device information, target domain knowledge, and contextual information.

[0333] In one embodiment, user information includes one or more of user profiles, user preferences, user behavior patterns, and historical interaction records; device information includes one or more of hardware configuration, software configuration, performance parameters, and device operating status; target domain knowledge includes one or more of concept definitions, entity relationships, rule logic, and domain models; and context information includes one or more of environmental state, situational information, and time information.

[0334] In one embodiment, the public semantic knowledge base includes one or more of general knowledge, domain expertise, device interaction knowledge, and a model library.

[0335] In one embodiment, general knowledge includes one or more of common sense, conceptual definitions, and general rules; domain expertise includes one or more of technical terms, cases, regulations, and industry standards; device interaction knowledge includes one or more of device types, interfaces, communication protocols, and operating instructions; and the model library includes one or more of statistical models, machine learning models, and deep learning models.

[0336] In one embodiment, the device further includes:

[0337] The indicator transmission module is used to transmit the performance indicators of semantic feature extraction completed on the cloud to the sending end. The performance indicators of semantic feature extraction completed on the cloud are used to instruct the sending end to compare the performance indicators of semantic feature extraction completed locally with those of semantic feature extraction completed on the cloud, based on the available local resources and the resources required for semantic feature extraction, and to determine whether to perform semantic feature extraction locally or on the cloud.

[0338] In one embodiment, the performance metrics for offloading to the cloud to complete semantic feature extraction include cloud completion time, data transmission time and data transmission energy consumption for transmitting the data to be processed to the server;

[0339] The indicator transmission module is used to receive data volume information of the data to be processed from the sending end; based on the data volume information, it calculates the cloud completion time, data transmission time and data transmission energy consumption, and transmits the cloud completion time, data transmission time and data transmission energy consumption to the sending end.

[0340] In one embodiment, the data volume information includes at least one of data size and data type.

[0341] Each module in the aforementioned semantic feature extraction device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0342] In one embodiment, a transmitting end is provided, taking a terminal device (hereinafter referred to as a terminal) as an example, as shown in Figure 15. Figure 15 is a schematic diagram of the structure of the terminal device provided in an embodiment of this application. The terminal device 700 shown in Figure 15 includes: at least one processor 701, a memory 707, at least one network interface 708, and a user interface 703. The various components in the terminal device 700 are coupled together through a bus system 705. It is understood that the bus system 705 is used to realize the connection and communication between these components. In addition to a data bus, the bus system 705 also includes a power bus, a control bus, and a status signal bus. However, for clarity, all buses are labeled as bus system 705 in Figure 15. In addition, this embodiment of the application also includes a transceiver 706, which may be multiple elements, including a transmitter and a receiver, providing a unit for communicating with various other devices over a transmission medium.

[0343] The user interface 703 may include a display, keyboard, or clicking device (e.g., mouse, trackball, touchpad, or touchscreen).

[0344] It is understood that the memory 707 in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DRRAM). The memory 707 of the systems and methods described in this application is intended to include, but is not limited to, these and any other suitable types of memory.

[0345] In some implementations, memory 707 stores elements such as executable modules or data structures, or subsets thereof, or extended sets thereof: operating system 7071 and application program 7072.

[0346] The operating system 7071 includes various system programs, such as a framework layer, a core library layer, and a driver layer, used to implement various basic business functions and handle hardware-based tasks. The application program 7072 includes various applications, such as a media player and a browser, used to implement various application functions. Programs implementing the methods of the embodiments of this application can be included in the application program 7072.

[0347] In this embodiment of the application, by calling the program or instructions stored in the memory 707, specifically the program or instructions stored in the application program 7072, the processor 701 is used to determine whether to perform semantic feature extraction locally or in the cloud based on the available local resources; if it is determined that semantic feature extraction will be performed locally, then semantic features will be extracted based on the local semantic knowledge base; and if it is determined that semantic feature extraction will be performed in the cloud, then the transmitter will be controlled to send instruction information to the server; the instruction information is used to instruct the server to extract semantic features based on the public semantic knowledge base.

[0348] The methods disclosed in some or all of the above embodiments of this application can also be applied to processor 701, or implemented by processor 701, or implemented by processor 701 in conjunction with other components (e.g., transceivers). Processor 701 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above methods can be completed by the integrated logic circuit of the hardware in processor 701 or by instructions in the form of software. The processor 701 mentioned above may be a general-purpose processor 701, a digital signal processor 701 (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor 701 may be a microprocessor 701, or it may be any conventional processor 701, etc. The steps of the method disclosed in the embodiments of this application can be directly implemented by the hardware decoding processor 701, or implemented by a combination of hardware and software modules in the decoding processor 701. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory 707, and the processor 701 reads the information in memory 707 and, in conjunction with its hardware, completes the steps of the above method.

[0349] It is understood that the embodiments described in this application can be implemented using hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors 701, controllers, microcontrollers, microprocessors 701, other electronic units for performing the functions described in this application, or combinations thereof.

[0350] For software implementation, the technology described in the embodiments of this application can be implemented by modules (e.g., procedures, functions, etc.) that perform the functions described in the embodiments of this application. The software code can be stored in memory and executed by processor 701. The memory can be implemented in processor 701 or external to processor 701.

[0351] In one embodiment, the processor 701 is specifically configured to determine whether the local system has the capability to complete semantic feature extraction based on the available local resources and the resources required for semantic feature extraction; if the local system has the capability to complete semantic feature extraction, it compares the performance metrics of completing semantic feature extraction locally with the performance metrics of completing semantic feature extraction offloaded to the cloud, and determines whether to perform semantic feature extraction locally or in the cloud based on the comparison result.

[0352] In one embodiment, the processor 701 is specifically used to evaluate the data to be processed and obtain the data volume information of the data to be processed; based on the data volume information, evaluate the resource requirements for running the local semantic feature extraction algorithm and use the resource requirements as the resources required for semantic feature extraction; if the available local resources are greater than the resources required for semantic feature extraction, then it is determined that the local processor has the ability to complete semantic feature extraction.

[0353] In one embodiment, the data volume information includes at least one of data size and data type; the resource requirements include at least one of computing requirements and memory requirements.

[0354] In one embodiment, the resources required for semantic feature extraction include semantic feature extraction computational requirements and semantic feature extraction memory requirements;

[0355] The processor 701 is specifically used to obtain locally available resources; locally available resources include locally available computing resources and locally available memory space; if the locally available computing resources are greater than the semantic feature extraction computing requirements and the locally available memory space is greater than the semantic feature extraction memory requirements, then it is determined that the local processor has the ability to complete semantic feature extraction.

[0356] In one embodiment, the performance metrics for semantic feature extraction performed locally include local completion time and local completion energy consumption; the performance metrics for semantic feature extraction performed offloaded to the cloud include cloud completion time, data transmission time and data transmission energy consumption for transmitting the data to be processed to the server.

[0357] The processor 701 is specifically used to control the transmitter to transmit data volume information to the server; the data volume information is used to instruct the server to calculate and return the cloud completion time, data transmission time, and data transmission energy consumption; if the local completion time is less than the sum of the data transmission time and the cloud completion time, and the local completion energy consumption is less than the data transmission energy consumption, then it is determined that semantic feature extraction is performed locally; if the local completion time is greater than or equal to the sum of the data transmission time and the cloud completion time, and / or the local completion energy consumption is greater than or equal to the data transmission energy consumption, then it is determined that semantic feature extraction is performed in the cloud.

[0358] In one embodiment, the processor 701 is specifically configured to extract semantic features based on a local semantic knowledge base; and to receive optimization suggestions from a server via a receiver, and optimize the semantic features according to the optimization suggestions.

[0359] In one embodiment, the processor 701 is further configured to create prompt words based on semantic communication goals and semantic features; control the transmitter to transmit the prompt words and necessary data to the server, so as to instruct the server to generate optimization suggestions for semantic features based on a public semantic knowledge base, and transmit the optimization suggestions to the sender.

[0360] In one embodiment, the necessary data includes one or more of the following: semantic features, raw data, data type, data format, data creation time, and data version.

[0361] In one embodiment, the cue word includes one or more of a description of the communication goal, a summary of semantic features, and the sender's intent.

[0362] In one embodiment, the cue words are presented in one or more of the following formats: text, vector, and structured data.

[0363] In one embodiment, semantic features include one or more of entity recognition, sentiment analysis, topic classification, and semantic role labeling.

[0364] In one embodiment, optimization suggestions include one or more of maintaining, expanding, reducing, merging, splitting, refining, and simplifying.

[0365] In one embodiment, the local semantic knowledge base includes one or more of user information, device information, target domain knowledge, and contextual information.

[0366] In one embodiment, user information includes one or more of user profiles, user preferences, user behavior patterns, and historical interaction records; device information includes one or more of hardware configuration, software configuration, performance parameters, and device operating status; target domain knowledge includes one or more of concept definitions, entity relationships, rule logic, and domain models; and context information includes one or more of environmental state, situational information, and time information.

[0367] In one embodiment, the public semantic knowledge base includes one or more of general knowledge, domain expertise, device interaction knowledge, and a model library.

[0368] In one embodiment, general knowledge includes one or more of common sense, conceptual definitions, and general rules; domain expertise includes one or more of technical terms, cases, regulations, and industry standards; device interaction knowledge includes one or more of device types, interfaces, communication protocols, and operating instructions; and the model library includes one or more of statistical models, machine learning models, and deep learning models.

[0369] In one embodiment, the instruction information includes a prompt word and necessary data.

[0370] In one embodiment, the processor 701 is specifically configured to create prompt words based on semantic communication targets and a local semantic knowledge base; control the transmitter to transmit the prompt words and necessary data to the server, so as to instruct the server to extract semantic features based on a public semantic knowledge base and transmit the semantic features to the sending end.

[0371] In one embodiment, the necessary data includes at least one of the following: raw data, data type, data format, data creation time, and data version.

[0372] In one embodiment, the prompt word includes one or more of the following: a description of the communication goal, a description of the communication scenario, the sender's intent, and a personalized tag.

[0373] In one embodiment, the cue words are presented in one or more of the following formats: text, vector, and structured data.

[0374] Figure 16 is a schematic diagram of the server structure provided in an embodiment of this application. The server 800 shown in Figure 16 includes at least one processor 801, a memory 803, and at least one network interface 805. The various components in the server 800 are coupled together via a bus system 807. It is understood that the bus system 807 is used to implement communication between these components. In addition to a data bus, the bus system 807 also includes a power bus, a control bus, and a status signal bus. However, for clarity, all buses are labeled as bus system 807 in Figure 16. Furthermore, this embodiment of the application also includes a transceiver 809, which may consist of multiple elements, including a transmitter and a receiver, providing a unit for communicating with various other devices over a transmission medium.

[0375] It is understood that the memory 803 in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DRRAM). The memory 803 of the systems and methods described in the embodiments of this application is intended to include, but is not limited to, these and any other suitable types of memory.

[0376] In some implementations, memory 803 stores the following elements: executable modules or data structures, or subsets thereof, or extended sets thereof: operating system 8031. Operating system 8031 ​​includes various system programs, such as a framework layer, core library layer, driver layer, etc., used to implement various basic business functions and handle hardware-based tasks.

[0377] In this embodiment of the application, by calling the program or instructions stored in the memory 803, the receiver is used to receive indication information from the sender; the processor 801 is used to extract semantic features based on the indication information and a public semantic knowledge base.

[0378] The methods disclosed in some or all of the above embodiments of this application can also be applied to processor 801, or implemented by processor 801, or implemented by processor 801 in conjunction with other components (e.g., transceivers). Processor 801 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in processor 801 or by instructions in the form of software. The processor 801 mentioned above may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the methods disclosed in the embodiments of this application can be directly embodied as being executed by a hardware decoding processor, or being executed by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory 803, and processor 801 reads the information in memory 803 and, in conjunction with its hardware, completes the steps of the above method.

[0379] It is understood that the embodiments described in this application can be implemented using hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in this application, or combinations thereof.

[0380] For software implementation, the technology described in the embodiments of this application can be implemented by modules (e.g., procedures, functions, etc.) that perform the functions described in the embodiments of this application. The software code can be stored in memory and executed by processor 801. The memory can be implemented in processor 801 or external to processor 801.

[0381] In one embodiment, the indication information is sent by the sending end when it determines that semantic feature extraction will be performed in the cloud based on available local resources; the indication information includes prompt words and necessary data.

[0382] In one embodiment, the processor 801 is further configured to extract semantic features based on a public semantic knowledge base in response to receiving prompt words and necessary data;

[0383] The transmitter is used to transmit semantic features to the sender.

[0384] In one embodiment, the necessary data includes at least one of the following: raw data, data type, data format, data creation time, and data version.

[0385] In one embodiment, the prompt word includes one or more of the following: a description of the communication goal, a description of the communication scenario, the sender's intent, and a personalized tag.

[0386] In one embodiment, the cue words are presented in one or more of the following formats: text, vector, and structured data.

[0387] In one embodiment, the transmitter is further configured to send optimization suggestions to the sender; the optimization suggestions are used to instruct the sender to optimize the semantic features extracted based on the local semantic knowledge base.

[0388] In one embodiment, the receiver is also used to receive prompts and necessary data from the sender;

[0389] The processor 801 is also used to generate optimization suggestions for semantic features based on a public semantic knowledge base, according to prompt words and necessary data.

[0390] In one embodiment, the semantic features are extracted by the sender based on a local semantic knowledge base; the prompt words are created by the sender based on the semantic communication target and semantic features.

[0391] In one embodiment, semantic features include one or more of entity recognition, sentiment analysis, topic classification, and semantic role labeling.

[0392] In one embodiment, the necessary data includes one or more of the following: semantic features, raw data, data type, data format, data creation time, and data version.

[0393] In one embodiment, the cue word includes one or more of a description of the communication goal, a summary of semantic features, and the sender's intent.

[0394] In one embodiment, the cue words are presented in one or more of the following formats: text, vector, and structured data.

[0395] In one embodiment, optimization suggestions include one or more of maintaining, expanding, reducing, merging, splitting, refining, and simplifying.

[0396] In one embodiment, the local semantic knowledge base includes one or more of user information, device information, target domain knowledge, and contextual information.

[0397] In one embodiment, user information includes one or more of user profiles, user preferences, user behavior patterns, and historical interaction records; device information includes one or more of hardware configuration, software configuration, performance parameters, and device operating status; target domain knowledge includes one or more of concept definitions, entity relationships, rule logic, and domain models; and context information includes one or more of environmental state, situational information, and time information.

[0398] In one embodiment, the public semantic knowledge base includes one or more of general knowledge, domain expertise, device interaction knowledge, and a model library.

[0399] In one embodiment, general knowledge includes one or more of common sense, conceptual definitions, and general rules; domain expertise includes one or more of technical terms, cases, regulations, and industry standards; device interaction knowledge includes one or more of device types, interfaces, communication protocols, and operating instructions; and the model library includes one or more of statistical models, machine learning models, and deep learning models.

[0400] In one embodiment, the transmitter is further configured to transmit performance metrics for semantic feature extraction offloaded to the cloud to the sending end; the performance metrics for semantic feature extraction offloaded to the cloud are used to instruct the sending end, when it is determined that it has the capability to complete semantic feature extraction locally based on available local resources and resources required for semantic feature extraction, to compare the performance metrics for semantic feature extraction performed locally with those for semantic feature extraction offloaded to the cloud, and to determine whether to perform semantic feature extraction locally or in the cloud based on the comparison result.

[0401] In one embodiment, the performance metrics for offloading to the cloud to complete semantic feature extraction include the cloud completion time, the data transmission time and data transmission energy consumption for transmitting the data to be processed to the server;

[0402] The receiver is also used to receive information about the amount of data to be processed from the sender;

[0403] The processor 801 is specifically used to calculate the cloud completion time, data transmission time, and data transmission energy consumption based on the data volume information;

[0404] The transmitter is also used to transmit cloud completion time, data transmission time, and data transmission energy consumption to the sending end.

[0405] In one embodiment, the data volume information includes at least one of data size and data type.

[0406] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0407] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0408] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0409] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0410] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0411] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A semantic feature extraction method, applied at a sending end, the method comprising: Based on available local resources, determine whether to perform semantic feature extraction locally or in the cloud; If it is determined that semantic feature extraction will be performed locally, then semantic features will be extracted based on the local semantic knowledge base; If it is determined that semantic feature extraction will be performed in the cloud, then an instruction message is sent to the server; the instruction message is used to instruct the server to extract semantic features based on a public semantic knowledge base.

2. The method according to claim 1, wherein, The step of determining whether to perform semantic feature extraction locally or in the cloud based on available local resources includes: Based on the available local resources and the resources required for semantic feature extraction, confirm whether the local system has the capability to complete semantic feature extraction. If the local machine has the capability to perform semantic feature extraction, then the performance metrics of performing semantic feature extraction locally and performing semantic feature extraction offloaded to the cloud are compared, and based on the comparison results, it is determined whether to perform semantic feature extraction locally or in the cloud.

3. The method according to claim 1, wherein, If it is determined that semantic feature extraction will be performed locally, then semantic features will be extracted based on the local semantic knowledge base, including: Based on the local semantic knowledge base, semantic features are extracted; Receive optimization suggestions from the server, and optimize the semantic features according to the optimization suggestions.

4. The method according to claim 3, wherein, Before the step of receiving optimization suggestions from the server and optimizing the semantic features according to the optimization suggestions, the method further includes: Based on the semantic communication objective and the semantic features, prompt words are created; The prompt words and necessary data are transmitted to the server to instruct the server to generate optimization suggestions for the semantic features based on the public semantic knowledge base, and then transmit the optimization suggestions to the sending end.

5. The method according to claim 1, wherein, The instruction information includes prompts and necessary data.

6. The method according to claim 5, wherein, If it is determined that semantic feature extraction will be performed in the cloud, then sending an instruction message to the server includes: The prompt words are created based on the semantic communication objective and the local semantic knowledge base; The prompt word and the necessary data are transmitted to the server, instructing the server to extract semantic features based on the public semantic knowledge base, and then transmit the semantic features to the sending end.

7. The method according to claim 2, wherein, The step of confirming whether the local system has the capability to complete semantic feature extraction based on the available local resources and the resources required for semantic feature extraction includes: The data to be processed is evaluated, and the data volume information of the data to be processed is obtained; Based on the data volume information, the resource requirements for running the local semantic feature extraction algorithm are assessed, and the resource requirements are taken as the resources required for the semantic feature extraction. If the available local resources are greater than the resources required for semantic feature extraction, then it is determined that the local system has the capability to complete semantic feature extraction.

8. The method according to claim 7, wherein, The data volume information includes at least one of data size and data type; the resource requirements include at least one of computing requirements and memory requirements.

9. The method according to claim 7, wherein, The resources required for semantic feature extraction include the computational requirements and memory requirements for semantic feature extraction. If the available local resources are greater than the resources required for semantic feature extraction, then determining that the local system has the capability to complete semantic feature extraction includes: Obtain the locally available resources; the locally available resources include locally available computing resources and locally available memory space; If the available local computing resources are greater than the semantic feature extraction computing requirements, and the available local memory space is greater than the semantic feature extraction memory requirements, then it is determined that the local system has the capability to complete semantic feature extraction.

10. The method according to claim 7, wherein, The performance metrics for completing semantic feature extraction locally include local completion time and local completion energy consumption; the performance metrics for completing semantic feature extraction on the cloud include cloud completion time, data transmission time and data transmission energy consumption for transmitting the data to be processed to the server. The comparison of performance metrics for semantic feature extraction performed locally and for semantic feature extraction performed offloaded to the cloud, and the determination of whether to perform semantic feature extraction locally or in the cloud based on the comparison results, includes: The data volume information is transmitted to the server; the data volume information is used to instruct the server to calculate and report back the cloud completion time, the data transmission time, and the data transmission energy consumption. If the local completion time is less than the sum of the data transmission time and the cloud completion time, and the local completion energy consumption is less than the data transmission energy consumption, then it is determined that semantic feature extraction will be performed locally. If the local completion time is greater than or equal to the sum of the data transmission time and the cloud completion time, and / or the local completion energy consumption is greater than or equal to the data transmission energy consumption, then semantic feature extraction is determined to be performed in the cloud.

11. The method according to claim 1, wherein, The local semantic knowledge base includes one or more of the following: user information, device information, target domain knowledge, and contextual information.

12. The method according to claim 11, wherein, The user information includes one or more of user profiles, user preferences, user behavior patterns, and historical interaction records; the device information includes one or more of hardware configuration, software configuration, performance parameters, and device operating status; the target domain knowledge includes one or more of concept definitions, entity relationships, rule logic, and domain models; and the context information includes one or more of environmental state, situational information, and time information.

13. The method according to claim 1, wherein, The public semantic knowledge base includes one or more of the following: general knowledge, domain-specific knowledge, device interaction knowledge, and model library.

14. The method according to claim 13, wherein, The general knowledge includes one or more of common sense, conceptual definitions, and general rules; the domain expertise includes one or more of professional terminology, cases, regulations, and industry standards; the device interaction knowledge includes one or more of device types, interfaces, communication protocols, and operating instructions; and the model library includes one or more of statistical models, machine learning models, and deep learning models.

15. The method according to claim 3, wherein, The semantic features include one or more of entity recognition, sentiment analysis, topic classification, and semantic role labeling.

16. The method according to claim 3, wherein, The optimization suggestions include one or more of the following: maintain, expand, reduce, merge, split, refine, and simplify.

17. The method according to claim 4, wherein, The necessary data includes one or more of the semantic features, raw data, data type, data format, data creation time, and data version.

18. The method according to claim 4, wherein, The prompt words include one or more of the following: a description of the communication target, a summary of semantic features, and the sender's intent.

19. The method according to claim 18, wherein, The prompts can be presented in one or more of the following formats: text, vector, and structured data.

20. The method according to claim 5, wherein, The necessary data includes at least one of the following: original data, data type, data format, data creation time, and data version.

21. The method according to claim 5, wherein, The prompt words include one or more of the following: a description of the communication target, a description of the communication scenario, the sender's intent, and a personalized tag.

22. The method according to claim 5, wherein, The prompts can be presented in one or more of the following formats: text, vector, and structured data.

23. The method according to claim 1, wherein, The step of determining whether to perform semantic feature extraction locally or in the cloud based on available local resources includes: Analyze and evaluate available channel resources; When it is determined, based on the results of channel resource assessment, that the current situation is not suitable for signal transmission, semantic feature extraction is performed locally.

24. A semantic feature extraction method, applied to a server, the method comprising: Receive indication information from the sender; Based on the indicated information, semantic features are extracted using a public semantic knowledge base.

25. The method according to claim 24, wherein, The instruction information is sent by the sending end when it determines that semantic feature extraction will be performed in the cloud based on available local resources; the instruction information includes prompt words and necessary data.

26. The method of claim 25, wherein, The step of extracting semantic features based on a public semantic knowledge base according to the indicated information includes: In response to receiving the prompt word and the necessary data, semantic features are extracted based on the public semantic knowledge base; The semantic features are transmitted to the sending end.

27. The method according to claim 24, wherein, The sending end determines whether to perform semantic feature extraction locally based on available local resources, and then extracts semantic features based on the local semantic knowledge base.

28. The method of claim 27, further comprising: Send optimization suggestions to the sending end; The optimization suggestion is used to instruct the sending end to optimize the semantic features extracted based on the local semantic knowledge base.

29. The method according to claim 28, wherein, Before sending the optimization suggestions to the sending end, the method includes: Receive prompts and necessary data from the sending end; Based on the prompt words and the necessary data, optimization suggestions for the semantic features are generated using the public semantic knowledge base.

30. The method according to claim 29, wherein, The prompt word is created by the sending end based on the semantic communication goal and the semantic features.

31. The method of claim 27, further comprising: The performance metrics for semantic feature extraction performed on the cloud are transmitted to the sending end. The performance metrics for semantic feature extraction performed on the cloud are used to instruct the sending end, when it is determined that the local end has the capability to perform semantic feature extraction based on the available local resources and the resources required for semantic feature extraction, to compare the performance metrics for semantic feature extraction performed locally with those for semantic feature extraction performed on the cloud, and to determine whether to perform semantic feature extraction locally or in the cloud based on the comparison result.

32. The method according to claim 31, wherein, The performance metrics for offloading to the cloud to complete semantic feature extraction include cloud completion time, data transmission time and energy consumption for transmitting the data to be processed to the server; the method further includes: Receive data volume information of the data to be processed from the sending end; Based on the data volume information, the cloud completion time, the data transmission time, and the data transmission energy consumption are calculated, and the cloud completion time, the data transmission time, and the data transmission energy consumption are transmitted to the sending end.

33. The method according to claim 32, wherein, The data volume information includes at least one of the data size and data type.

34. A semantic feature extraction device, applied at a sending end, the device comprising: The extraction method determination module is used to determine whether to perform semantic feature extraction locally or in the cloud, based on available local resources. The local extraction module is used to extract semantic features based on the local semantic knowledge base if it is determined that semantic feature extraction will be performed locally. The instruction module is used to send instruction information to the server if it is determined that semantic feature extraction will be performed in the cloud; the instruction information is used to instruct the server to extract semantic features based on a public semantic knowledge base.

35. A semantic feature extraction device, applied to a server, the device comprising: The information receiving module is used to receive indication information from the sending end; The feature extraction module is used to extract semantic features based on a public semantic knowledge base according to the indicated information.

36. A transmitter, comprising: Transmitter and processor; The processor is used to determine whether to perform semantic feature extraction locally or in the cloud, based on available local resources. If it is determined that semantic feature extraction will be performed locally, then semantic features will be extracted based on the local semantic knowledge base; And if it is determined that semantic feature extraction is performed in the cloud, the transmitter is controlled to send instruction information to the server; the instruction information is used to instruct the server to extract semantic features based on a public semantic knowledge base.

37. A server, comprising: Receiver and processor; The receiver is used to receive indication information from the sender; The processor is used to extract semantic features based on a public semantic knowledge base according to the instruction information.

38. A communication system comprising the transmitting end of claim 36 and the server of claim 37.

39. A computer-readable storage medium having a computer program stored thereon, wherein, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 33.

40. A computer program product comprising a computer program, wherein, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 33.

Images