CN118138192A - Semantic communication method, semantic communication device, electronic equipment and storage medium - Google Patents

Semantic communication method, semantic communication device, electronic equipment and storage medium Download PDF

Info

Publication number
CN118138192A
CN118138192A CN202410058360.0A CN202410058360A CN118138192A CN 118138192 A CN118138192 A CN 118138192A CN 202410058360 A CN202410058360 A CN 202410058360A CN 118138192 A CN118138192 A CN 118138192A
Authority
CN
China
Prior art keywords
semantic
answer
adjustment
semantic information
text data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202410058360.0A
Other languages
Chinese (zh)
Inventor
罗涛
张建桐
张志龙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Posts and Telecommunications
Original Assignee
Beijing University of Posts and Telecommunications
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing University of Posts and Telecommunications filed Critical Beijing University of Posts and Telecommunications
Priority to CN202410058360.0A priority Critical patent/CN118138192A/en
Publication of CN118138192A publication Critical patent/CN118138192A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/0001—Systems modifying transmission characteristics according to link quality, e.g. power backoff
    • H04L1/0006—Systems modifying transmission characteristics according to link quality, e.g. power backoff by adapting the transmission format
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/0001—Systems modifying transmission characteristics according to link quality, e.g. power backoff
    • H04L1/0015—Systems modifying transmission characteristics according to link quality, e.g. power backoff characterised by the adaptation strategy

Landscapes

  • Engineering & Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Theoretical Computer Science (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Machine Translation (AREA)

Abstract

The application provides a semantic communication method, a device, electronic equipment and a storage medium, wherein a sending end generates a first transmission question and first adjustment semantic information which are related to text data to be transmitted based on the text data to be transmitted, determines a first target answer corresponding to the first transmission question and stores the first target answer, a receiving end determines a first answer corresponding to the first transmission question based on the first adjustment semantic information, then predicts semantic processing of the receiving end in a question-and-answer mode, adjusts the generated first adjustment semantic information at the sending end so as to achieve the purpose of generating second adjustment semantic information which can be more easily understood by the receiving end, and therefore the receiving end can generate first recovery text data consistent with the text data to be transmitted according to the received second adjustment semantic information, and semantic communication performance is guaranteed.

Description

Semantic communication method, semantic communication device, electronic equipment and storage medium
Technical Field
The present application relates to the field of communications technologies, and in particular, to a semantic communication method, a semantic communication device, an electronic device, and a storage medium.
Background
When the sending end and the receiving end of the semantic communication system have different semantic processing rules, the condition that the semantic understanding capability of the receiving end is reduced can occur, so that the performance of the semantic communication system is reduced, and normal communication can be possibly affected.
Therefore, when the sending end and the receiving end of the semantic communication system have different semantic processing rules, how to effectively perform semantic communication is easier for the receiving end to understand to be the technical problem to be solved.
Disclosure of Invention
In view of the above, the present application aims to provide a semantic communication method, a semantic communication device, an electronic device and a storage medium for solving the above technical problems.
Based on the above object, a first aspect of the present application provides a semantic communication method, applied to a semantic communication system, the system including a transmitting end and a receiving end, the method comprising:
The sending end obtains text data to be transmitted, generates a first transmission question and first adjustment semantic information associated with the text data to be transmitted based on the text data to be transmitted, determines a first target answer corresponding to the first transmission question, stores the first target answer, and transmits the first adjustment semantic information and the first transmission question to the receiving end;
the receiving end receives first adjustment semantic information and a first transmission problem transmitted by the transmitting end, determines a first reply answer corresponding to the first transmission problem based on the first adjustment semantic information, and transmits the first reply answer to the transmitting end;
The sending terminal receives a first answer transmitted by the receiving terminal, determines first semantic similarity between the first answer and the first target answer, and adjusts the first adjustment semantic information according to the first semantic similarity to generate second adjustment semantic information;
The receiving end receives second adjustment semantic information transmitted by the transmitting end, performs semantic processing according to the second adjustment semantic information, and generates first recovery text data, wherein the first recovery text data is identical to the text data to be transmitted.
Optionally, the generating, based on the text data to be transmitted, a first transmission problem and first adjustment semantic information associated with the text data to be transmitted includes:
the sending end inputs the text data to be transmitted into a pre-trained problem generation model, and generates a first transmission problem associated with the text data to be transmitted through the problem generation model;
The sending end extracts initial semantic information from text data to be transmitted according to a preset first semantic processing rule, continuously adjusts the initial semantic information to obtain adjusted initial semantic information, and sends the adjusted initial semantic information to the receiving end in the continuous adjustment process so that the receiving end can determine a second answer corresponding to the first transmission problem according to the adjusted initial semantic information and send the second answer back to the sending end;
The sending end determines second semantic similarity between the second answer and the first target answer until the value of the second semantic similarity is the same as the value of the initial semantic similarity determined randomly, and takes the adjusted initial semantic information corresponding to the second semantic similarity as first adjustment semantic information.
Optionally, the determining a first target answer corresponding to the first transmission question includes:
The sending end inputs the text data to be transmitted and the first transmission problem into a pre-trained answer generation model, and a plurality of answers corresponding to the first transmission problem are generated through the answer generation model;
and the sending end selects a first target answer from the multiple answers according to a preset question selection vector.
Optionally, the determining, based on the first adjustment semantic information, a first answer to a response corresponding to the first transmission question includes:
The receiving end performs recovery processing according to the first adjustment semantic information by using a preset second semantic processing rule to generate second recovery text data, wherein the second semantic processing rule represents a transformation rule between the first adjustment semantic information and the second recovery text data;
And the receiving end replies the first transmission problem by utilizing the second recovery text data according to a preset reply rule, and generates a first reply answer corresponding to the first transmission problem, wherein the reply rule is a rule for calculating the first reply answer corresponding to the first transmission problem according to the specification of the second recovery text data.
Optionally, the determining the first semantic similarity between the first answer and the first target answer includes:
The sending end encodes the first answer and the first target answer respectively, and generates an encoded answer corresponding to the first answer and an encoded first target answer corresponding to the first target answer;
The sending end performs product processing on the coded reply answer and the coded first target answer to generate a product processing result;
And the sending end acquires the value of the question selection vector corresponding to the first target answer, and carries out ratio processing on the product processing result and the value of the question selection vector corresponding to the first target answer to obtain the first semantic similarity between the first answer and the first target answer.
Optionally, the adjusting the first adjustment semantic information according to the first semantic similarity generates second adjustment semantic information, including:
The sending end obtains a historical first semantic similarity value corresponding to a previous adjusting process corresponding to a current adjusting process, obtains a reward value corresponding to the historical first semantic similarity value, adjusts first adjusting semantic information by using the reward value to obtain adjusted first adjusting semantic information, sends the adjusted first adjusting semantic information to the receiving end, so that the receiving end determines a third answer corresponding to the first transmission problem according to the adjusted first adjusting semantic information, and returns the third answer to the sending end;
The sending end determines the first semantic similarity between the third answer and the first target answer until the preset first adjustment process times are reached, a plurality of first semantic similarities are obtained, the value of the largest first semantic similarity is selected from the values of the plurality of first semantic similarities, and the adjusted semantic information corresponding to the value of the largest first semantic similarity is used as second adjustment semantic information.
Optionally, the adjusting the first adjustment semantic information according to the first semantic similarity generates second adjustment semantic information, including:
The sending end obtains a preset question selection vector, randomly adjusts the question selection vector to obtain an adjusted question selection vector, determines a second transmission question corresponding to the adjusted question selection vector, and a second target answer corresponding to the second transmission question, stores the second target answer, and transmits the second transmission question and the first adjustment semantic information to the receiving end so that the receiving end can determine a fourth answer corresponding to the second transmission question according to the first adjustment semantic information;
The sending end determines third semantic similarity between the fourth answer and the first target answer until the preset second adjustment process times are reached, a plurality of third semantic similarities are obtained, the value of the smallest third semantic similarity is selected from the plurality of third semantic similarities, an adjusted question selection vector corresponding to the value of the smallest third semantic similarity is used as a target question selection vector, and a target transmission question corresponding to the target question selection vector is obtained;
The sending end obtains a historical first semantic similarity value corresponding to a previous adjusting process corresponding to a current adjusting process, obtains a rewarding value corresponding to the historical first semantic similarity value, adjusts first adjusting semantic information by utilizing the rewarding value to obtain adjusted first adjusting semantic information, sends the adjusted first adjusting semantic information to the receiving end, so that the receiving end determines a fourth answer corresponding to the first transmission problem according to the adjusted first adjusting semantic information, and returns the fourth answer to the sending end;
The sending end determines the first semantic similarity between the fourth answer and the first target answer until the preset third adjustment process times are reached, a plurality of first semantic similarities are obtained, the value of the largest first semantic similarity is selected from the values of the plurality of first semantic similarities, and the adjusted semantic information corresponding to the value of the largest first semantic similarity is used as second adjustment semantic information.
Based on the same inventive concept, a second aspect of the present application provides a semantic communication apparatus, the apparatus being provided in a semantic communication system, the system including a transmitting end and a receiving end, the apparatus comprising:
The sending end is configured to acquire text data to be transmitted, generate a first transmission problem and first adjustment semantic information associated with the text data to be transmitted based on the text data to be transmitted, determine a first target answer corresponding to the first transmission problem, store the first target answer, and transmit the first adjustment semantic information and the first transmission problem to the receiving end;
The receiving terminal is configured to receive first adjustment semantic information and a first transmission question transmitted by the transmitting terminal, determine a first answer corresponding to the first transmission question based on the first adjustment semantic information, and transmit the first answer to the transmitting terminal;
The sending end is configured to receive a first answer transmitted by the receiving end, determine a first semantic similarity between the first answer and the first target answer, and adjust the first adjustment semantic information according to the first semantic similarity to generate second adjustment semantic information;
The receiving end is configured to receive the second adjustment semantic information transmitted by the transmitting end, perform semantic processing according to the second adjustment semantic information, and generate first recovery text data, wherein the first recovery text data is the same as the text data to be transmitted.
Based on the same inventive concept, a third aspect of the present application provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable by the processor, the processor implementing the method as described in the first aspect above when executing the computer program.
Based on the same inventive concept, a fourth aspect of the present application provides a non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the method of the first aspect above.
From the above, it can be seen that, according to the semantic communication method, apparatus, electronic device and storage medium provided by the present application, the sending end generates the first transmission problem associated with the text data to be transmitted and the first adjustment semantic information based on the text data to be transmitted, determines the first target answer corresponding to the first transmission problem and stores the first target answer, and the receiving end determines the first answer corresponding to the first transmission problem based on the first adjustment semantic information, so as to estimate the semantic processing of the receiving end in the form of question-answer, and adjusts the generated first adjustment semantic information at the sending end, so as to achieve the purpose of generating the second adjustment semantic information that can be understood by the receiving end more easily, thus the receiving end performs semantic processing according to the received second adjustment semantic information, and can generate the first recovery text data consistent with the text data to be transmitted, thereby guaranteeing the semantic communication performance.
Drawings
In order to more clearly illustrate the technical solutions of the present application or related art, the drawings that are required to be used in the description of the embodiments or related art will be briefly described below, and it is apparent that the drawings in the following description are only embodiments of the present application, and other drawings may be obtained according to the drawings without inventive effort to those of ordinary skill in the art.
FIG. 1 is a flow chart of a semantic communication method according to an embodiment of the present application;
FIG. 2 is a schematic diagram of a semantic communication flow according to an embodiment of the present application;
FIG. 3 is a schematic structural diagram of a semantic communication device according to an embodiment of the present application;
Fig. 4 is a schematic diagram of an electronic device according to an embodiment of the application.
Detailed Description
The present application will be further described in detail below with reference to specific embodiments and with reference to the accompanying drawings, in order to make the objects, technical solutions and advantages of the present application more apparent.
It should be noted that unless otherwise defined, technical or scientific terms used in the embodiments of the present application should be given the ordinary meaning as understood by one of ordinary skill in the art to which the present application belongs. The terms "first," "second," and the like, as used in embodiments of the present application, do not denote any order, quantity, or importance, but rather are used to distinguish one element from another. The word "comprising" or "comprises", and the like, means that elements or items preceding the word are included in the element or item listed after the word and equivalents thereof, but does not exclude other elements or items. The terms "connected" or "connected," and the like, are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "upper", "lower", "left", "right", etc. are used merely to indicate relative positional relationships, which may also be changed when the absolute position of the object to be described is changed.
It will be appreciated that before using the technical solutions of the embodiments of the present application, the user is informed of the type, the range of use, the use scenario, etc. of the related personal information in an appropriate manner, and the authorization of the user is obtained.
For example, in response to receiving an active request from a user, a prompt is sent to the user to explicitly prompt the user that the operation it is requesting to perform will require personal information to be obtained and used with the user. Therefore, the user can select whether to provide personal information to the software or hardware such as the electronic equipment, the application program, the server or the storage medium for executing the operation of the technical scheme according to the prompt information.
As an alternative but non-limiting implementation, in response to receiving an active request from a user, the manner in which the prompt information is sent to the user may be, for example, a popup, in which the prompt information may be presented in a text manner. In addition, a selection control for the user to select to provide personal information to the electronic device in a 'consent' or 'disagreement' manner can be carried in the popup window.
It will be appreciated that the above-described notification and user authorization acquisition process is merely illustrative, and not limiting of the implementation of the present application, and that other ways of satisfying relevant legal regulations may be applied to the implementation of the present application.
Embodiments of the present application are described in detail below with reference to the accompanying drawings.
Most of the existing semantic communication technologies assume that a receiving end and a sending end have the same semantic processing rule, so that the same semantic information can be correctly understood. Obviously, when the sending end and the receiving end have different semantic processing rules, the semantic understanding capability of the receiving end is reduced, the performance of the semantic communication system is reduced, and even abnormal communication is caused seriously.
Therefore, when the sending end and the receiving end of the semantic communication system have different semantic processing rules, how to effectively perform semantic communication is easier for the receiving end to understand to be the technical problem to be solved.
The embodiment of the application provides a semantic communication method, a sending end generates a first transmission question and first adjustment semantic information which are related to text data to be transmitted based on the text data to be transmitted, a first target answer corresponding to the first transmission question is determined and stored, a receiving end determines a first answer corresponding to the first transmission question based on the first adjustment semantic information, then the semantic processing of the receiving end is estimated in a question-and-answer mode, the sending end adjusts the generated first adjustment semantic information to achieve the purpose of generating second adjustment semantic information which can be understood by the receiving end more easily, so that the receiving end can generate first recovery text data consistent with the text data to be transmitted by carrying out semantic processing according to the received second adjustment semantic information, and the semantic communication performance is guaranteed.
As shown in fig. 1, the method of the present embodiment is applied to a semantic communication system, where the system includes a transmitting end and a receiving end, and the method includes:
step 101, the transmitting end obtains text data to be transmitted, generates a first transmission question and first adjustment semantic information associated with the text data to be transmitted based on the text data to be transmitted, determines a first target answer corresponding to the first transmission question, stores the first target answer, and transmits the first adjustment semantic information and the first transmission question to the receiving end.
In this step, the text data to be transmitted is a data set consisting of a series of words, typically describing the actual objects and their relationships, which can be analyzed, processed and mined to extract useful information therefrom.
An entity is a thing or concept with a unique, distinguishable identity in the text data to be transmitted. These things can be concrete objects, places, etc., or abstract concepts, events. Entities are usually physically present and may be abstract.
A relationship is a connection, association or interaction that exists between entities. Relationships describe the relevance between different entities, which can be direct physical connection or abstract semantic relationships.
A triplet is a structure commonly used in information presentation to describe a relationship between entities. Triplets can be generally expressed asWherein/>P n is the relationship between entities.
Semantic information refers to meaning, concepts or meanings conveyed in text data to be transmitted. Reference is made to words, phrases, sentences and the actual meaning expressed by larger units of text, not just to the superficial literal form. Semantic information is a deep understanding of language, including comprehensive consideration of vocabulary, grammar, and context.
In the present application, semantic information is composed of triples. For a piece of text data to be transmitted, the semantic information can be characterized asWhere G n represents the number of triples contained in the semantic information.
As shown in fig. 2, for text data to be transmittedThe sending end carries out estimation based on text data to be transmitted to generate a problem set/>First regulatory semantic information/>And selecting vector a k from the problem set/>, based on the determined problemIs to choose the first transmission problem/>Then using the text data to be transmitted, and the first transmission problem/>Generating standard answers(I.e., the first target answer).
Adjusting the semantic informationAnd first transmission problem/>Transmitting to the receiving end.
Step 102, the receiving end receives the first adjustment semantic information and the first transmission question transmitted by the transmitting end, determines a first answer corresponding to the first transmission question based on the first adjustment semantic information, and transmits the first answer to the transmitting end.
In this step, as shown in fig. 2, the receiving end adjusts the first semantic informationRestoring to text data/>And answering the first transmission question according to the restored text data to obtain a first answer/>And feeding back to the transmitting end.
Step 103, the sending end receives the first answer transmitted by the receiving end, determines a first semantic similarity between the first answer and the first target answer, and adjusts the first adjustment semantic information according to the first semantic similarity to generate second adjustment semantic information.
In this step, the receiving end user then feeds back the first answer to the transmitting end user. And comparing the difference between the standard answers and the answers of the receiving end users by the sending end users, estimating the knowledge of the receiving end users, and further optimizing the semantic extraction and question generation module.
As shown in FIG. 2, to ensure that the receiving end can correctly understand the semantic information, the receiving end compares the standard answersAnd the first answer/>The difference is a first semantic similarity, the first semantic similarity is used as a basis, semantic processing of the receiving end is estimated, and first adjustment semantic information generated by the sending end is adjusted to generate second adjustment semantic information which is easier to understand by the receiving end.
Step 104, the receiving end receives the second adjustment semantic information transmitted by the transmitting end, performs semantic processing according to the second adjustment semantic information, and generates first recovery text data, wherein the first recovery text data is the same as the text data to be transmitted.
In the step, the receiving end performs semantic processing according to the received second adjustment semantic information to generate first recovery text data consistent with the text data to be transmitted, so that the performance of a semantic communication system is ensured, and the situation of affecting normal communication is avoided.
According to the scheme, the sending end generates the first transmission problem and the first adjustment semantic information related to the text data to be transmitted based on the text data to be transmitted, determines the first target answer corresponding to the first transmission problem and stores the first target answer, and the receiving end determines the first reply answer corresponding to the first transmission problem based on the first adjustment semantic information, so that the semantic processing of the receiving end is estimated through a question-answer form, the first adjustment semantic information generated by the sending end is adjusted to achieve the purpose of generating the second adjustment semantic information which can be understood by the receiving end more easily, and therefore the receiving end can generate the first recovery text data consistent with the text data to be transmitted through semantic processing according to the received second adjustment semantic information, and semantic communication performance is guaranteed.
In some embodiments, in step 101, the generating, based on the text data to be transmitted, a first transmission problem associated with the text data to be transmitted and first adjustment semantic information includes:
And A1, inputting the text data to be transmitted into a pre-trained problem generation model by the transmitting end, and generating a first transmission problem associated with the text data to be transmitted through the problem generation model.
And step A2, the transmitting end extracts initial semantic information from text data to be transmitted according to a preset first semantic processing rule, continuously adjusts the initial semantic information to obtain adjusted initial semantic information, and transmits the adjusted initial semantic information to the receiving end in the continuous adjustment process so that the receiving end can determine a second answer corresponding to the first transmission problem according to the adjusted initial semantic information, and transmits the second answer back to the transmitting end.
And A3, the sending end determines a second semantic similarity between the second answer and the first target answer until the value of the second semantic similarity is the same as the value of the initial semantic similarity determined randomly, and takes the adjusted initial semantic information corresponding to the second semantic similarity as first adjustment semantic information.
In the above scheme, based on the text data to be transmittedTraining is performed in advance by using a neural network model to obtain a problem generation model capable of generating a transmission problem according to text data. And then generating a first transmission problem based on the text data to be transmitted by using the problem generation model.
Wherein the problem generation model is preferentially a large language model (Large Language Model, LLM).
Thereby making the generated first transmission problem more accurate.
For text data to be transmittedGenerating initial semantic information based on a first semantic processing ruleWhere k=1, 2, …, K is the number of text data to be transmitted, K is the total number of text data to be transmitted, and G k is the number of triples contained in the semantic information. Initial semantic information/>, due to the variability of semantic processing rules between the receiving and transmitting endsMay not be directly understood by the receiving end. Thus, the initial semantic information/>Tuning to first regulatory semantic information/>Wherein/>The number of triples contained for the first adjusted semantic information.
In some embodiments, in step 101, the determining a first target answer corresponding to the first transmission question includes:
And B1, the transmitting end inputs the text data to be transmitted and the first transmission questions into a pre-trained answer generation model, and a plurality of answers corresponding to the first transmission questions are generated through the answer generation model.
And step B2, the transmitting end selects a first target answer from the plurality of answers according to a preset question selection vector.
In the above scheme, based on the text data to be transmittedTraining is performed in advance by using a neural network model to obtain an answer generation model capable of generating an answer according to text data and transmission questions. And generating a plurality of answers based on the text data to be transmitted and the first transmission question by utilizing the answer generation model.
Wherein the problem generation model is preferentially a large language model (Large Language Model, LLM).
Thus, the generated multiple answers are more accurate.
The answers to the questions are stored in the form of a set of questions, i.eWherein/>For the text data/>, to be transmittedQ k is the total number of the Q-th question to generate the first transmission question.
Because of limited spectrum resources, the communication system cannot transmit all the first transmission problems to the receiving end user, and only a part of the first transmission problems can be selected for transmission.
Let the problem selection vector beWherein/>To indicate whether or not to transmit a problemIf/>The problem/>, is transmitted The problem/>, is not transmittedThe selected set of questions (i.e., the first transmission question) may be represented as follows:
the corresponding first target answer is represented as follows:
Wherein, Is a problem/>And a corresponding first target answer. Finally, the selected question set/>And first regulatory semantic information/>Transmitting the sent end to the receiving end.
In some embodiments, in step 102, the determining, based on the first adjustment semantic information, a first answer to a reply corresponding to the first transmission question includes:
And C1, the receiving end performs recovery processing according to the first regulation semantic information by using a preset second semantic processing rule to generate second recovery text data, wherein the second semantic processing rule represents a transformation rule between the first regulation semantic information and the second recovery text data.
And C2, the receiving end replies the first transmission problem by utilizing the second recovery text data according to a preset reply rule, and a first reply answer corresponding to the first transmission problem is generated, wherein the reply rule is a rule for calculating the first reply answer corresponding to the first transmission problem according to the specification of the second recovery text data.
In the above scheme, according to the received first adjustment semantic informationThe first adjusting semantic information/>, by using the second semantic processing rule of the receiving endPerforming restoration to generate second restored text data/>
To judge whether the receiving end correctly understands the first adjustment semantic informationThe receiving end needs to recover the text data/>, according to the generated secondFor the first transmission problem/>The answer is made.
First transmission problem of receiving endThe answer (i.e., the first answer) is represented as follows:
Wherein, Is the receiving end concerning the problem/>Is a response to a request for a response from the host computer. /(I)Is transmitted back to the transmitting end to estimate the difference of semantic processing rules between the receiving end and the transmitting end.
In some embodiments, in step 103, the determining a first semantic similarity between the first answer to answer and the first target answer includes:
And D1, the transmitting end respectively codes the first answer and the first target answer, and generates a coded answer corresponding to the first answer and a coded first target answer corresponding to the first target answer.
And D2, the sending end performs product processing on the coded reply answer and the coded first target answer to generate a product processing result.
And D3, the sending end obtains the value of the question selection vector corresponding to the first target answer, and carries out ratio processing on the product processing result and the value of the question selection vector corresponding to the first target answer to obtain the first semantic similarity between the first answer and the first target answer.
In the above scheme, based on the first target answer and the first answer, by measuring the semantic similarity between the first target answer and the first answer, the difference of the semantic processing rules between the sending end and the receiving end is estimated to guide the first adjustment semantic informationAnd the generation of the question selection vector a k.
Encoding the first target answer into an encoded first target answer using a depth neural network based encoder pi (·)Encoding the first reply answer fed back by the receiving end into an encoded reply answer
The first semantic similarity between the first target answer and the first reply answer is expressed as follows:
Wherein, Values representing problem selection vectors,/>Representing the result of the product processing,/>Representing a first semantic similarity.
In some embodiments, in step 103, the adjusting the first adjustment semantic information according to the first semantic similarity, to generate second adjustment semantic information includes:
And E1, the sending end acquires a value of historical first semantic similarity corresponding to a previous adjusting process corresponding to a current adjusting process, acquires a reward value corresponding to the value of the historical first semantic similarity, adjusts first adjusting semantic information by utilizing the reward value to obtain adjusted first adjusting semantic information, sends the adjusted first adjusting semantic information to the receiving end, so that the receiving end determines a third answer corresponding to the first transmission problem according to the adjusted first adjusting semantic information, and returns the third answer to the sending end.
And E2, the sending end determines the first semantic similarity between the third answer and the first target answer until the preset first adjustment process times are reached, a plurality of first semantic similarities are obtained, the value of the largest first semantic similarity is selected from the values of the plurality of first semantic similarities, and the adjusted semantic information corresponding to the value of the largest first semantic similarity is used as second adjustment semantic information.
In the above scheme, the receiving end determines the third answer corresponding to the first transmission problem according to the first adjustment semantic information through iteration adjustment of the first adjustment semantic information, determines the first semantic similarity between the third answer and the first target answer, reaches the preset first adjustment process times, obtains a plurality of first semantic similarities, selects the value of the largest first semantic similarity from the values of the plurality of first semantic similarities, takes the adjusted first adjustment semantic information corresponding to the value of the largest first semantic similarity as the second adjustment semantic information, and at the moment, the receiving end can correctly understand the second adjustment semantic information, and further can recover the recovered text data consistent with the text data to be transmitted of the sending end, thereby guaranteeing the performance of the semantic communication system and avoiding the situation of affecting normal communication.
In some embodiments, in step 103, the adjusting the first adjustment semantic information according to the first semantic similarity, to generate second adjustment semantic information includes:
Step F1, the sending end obtains a preset question selection vector, randomly adjusts the question selection vector to obtain an adjusted question selection vector, determines a second transmission question corresponding to the adjusted question selection vector, and a second target answer corresponding to the second transmission question and stores the second target answer, and transmits the second transmission question and the first adjustment semantic information to the receiving end so that the receiving end can determine a fourth answer corresponding to the second transmission question according to the first adjustment semantic information.
And F2, the sending end determines the third semantic similarity between the fourth answer and the first target answer until the preset second adjusting process times are reached, a plurality of third semantic similarities are obtained, the value of the smallest third semantic similarity is selected from the plurality of third semantic similarities, an adjusted question selection vector corresponding to the value of the smallest third semantic similarity is used as a target question selection vector, and a target transmission question corresponding to the target question selection vector is acquired.
And F3, the sending end obtains a value of historical first semantic similarity corresponding to a last adjusting process corresponding to the current adjusting process, obtains a reward value corresponding to the value of the historical first semantic similarity, adjusts the first adjusting semantic information by utilizing the reward value to obtain adjusted first adjusting semantic information, sends the adjusted first adjusting semantic information to the receiving end, so that the receiving end determines a fourth answer corresponding to the first transmission problem according to the adjusted first adjusting semantic information, and returns the fourth answer to the sending end.
And F4, the sending end determines the first semantic similarity between the fourth answer and the first target answer until the preset third adjustment process times are reached, a plurality of first semantic similarities are obtained, the value of the largest first semantic similarity is selected from the values of the plurality of first semantic similarities, and the adjusted semantic information corresponding to the value of the largest first semantic similarity is used as second adjustment semantic information.
In the above scheme, the question selection vector is iteratively adjusted and transmitted to a second transmission question which is difficult to understand by the receiving end, so that the receiving end determines a second target answer corresponding to the second transmission question, and the sending end determines a third semantic similarity between the fourth answer and the first target answer until the preset second adjustment process times are reached, so as to obtain a plurality of third semantic similarities, and the minimum value of the third semantic similarity is selected from the plurality of third semantic similarities.
And taking the adjusted question selection vector corresponding to the minimum value of the third semantic similarity as a target question selection vector, and acquiring a target transmission question corresponding to the target question selection vector.
The first adjustment semantic information is adjusted through iteration, so that the receiving end determines a fourth answer corresponding to the target transmission problem according to the first adjustment semantic information, determines first semantic similarity between the fourth answer and the first target answer until the preset third adjustment process times are reached, obtains a plurality of first semantic similarities, selects the value of the largest first semantic similarity from the values of the plurality of first semantic similarities, and takes the adjusted first adjustment semantic information corresponding to the largest first semantic similarity as second adjustment semantic information.
The problem selection vector is adjusted, the value of the minimum third semantic similarity is determined, the value of the maximum first semantic similarity is determined, and countermeasures adjustment is formed through the value of the minimum third semantic similarity and the value of the maximum first semantic similarity, so that the receiving end can correctly understand the second adjustment semantic information, further recover the recovered text data consistent with the text data to be transmitted of the sending end, the performance of a semantic communication system is guaranteed, and the situation that normal communication is affected is avoided.
It should be noted that, the method of the embodiment of the present application may be performed by a single device, for example, a computer or a server. The method of the embodiment can also be applied to a distributed scene, and is completed by mutually matching a plurality of devices. In the case of such a distributed scenario, one of the devices may perform only one or more steps of the method of an embodiment of the present application, the devices interacting with each other to accomplish the method.
It should be noted that the foregoing describes some embodiments of the present application. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments described above and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
Based on the same inventive concept, the application also provides a semantic communication device corresponding to the method of any embodiment.
Referring to fig. 3, the semantic communication apparatus is disposed in a semantic communication system, the system includes a transmitting end and a receiving end, and the apparatus includes:
The transmitting end 301 is configured to obtain text data to be transmitted, generate a first transmission question and first adjustment semantic information associated with the text data to be transmitted based on the text data to be transmitted, determine a first target answer corresponding to the first transmission question, store the first target answer, and transmit the first adjustment semantic information and the first transmission question to the receiving end;
The receiving end 302 is configured to receive the first adjustment semantic information and the first transmission question transmitted by the sending end, determine a first answer corresponding to the first transmission question based on the first adjustment semantic information, and transmit the first answer to the sending end;
The sending end 301 is configured to receive a first answer transmitted by the receiving end, determine a first semantic similarity between the first answer and the first target answer, and adjust the first adjustment semantic information according to the first semantic similarity to generate second adjustment semantic information;
The receiving end 302 is configured to receive the second adjustment semantic information transmitted by the transmitting end, perform semantic processing according to the second adjustment semantic information, and generate first recovered text data, where the first recovered text data is the same as the text data to be transmitted.
In some embodiments, the transmitting end 301 is configured to:
The transmitting end 301 is configured to input the text data to be transmitted into a pre-trained question generation model, and generate a first transmission question associated with the text data to be transmitted through the question generation model;
The transmitting end 301 is configured to extract initial semantic information from text data to be transmitted according to a preset first semantic processing rule, continuously adjust the initial semantic information to obtain adjusted initial semantic information, and send the adjusted initial semantic information to the receiving end in a continuous adjustment process, so that the receiving end determines a second answer corresponding to the first transmission problem according to the adjusted initial semantic information, and returns the second answer to the transmitting end;
the sending end 301 is configured to determine a second semantic similarity between the second answer and the first target answer, until the value of the second semantic similarity is the same as the value of the initial semantic similarity determined randomly, and take the adjusted initial semantic information corresponding to the second semantic similarity as first adjustment semantic information.
In some embodiments, the transmitting end 301 is configured to:
The transmitting end 301 is configured to input the text data to be transmitted and the first transmission question into a pre-trained answer generation model, and generate a plurality of answers corresponding to the first transmission question through the answer generation model;
the transmitting end 301 is configured to select a first target answer from the multiple answers according to a preset question selection vector.
In some embodiments, the receiving end 302 is configured to:
the receiving end 302 is configured to perform recovery processing according to the first adjustment semantic information by using a preset second semantic processing rule, and generate second recovery text data, where the second semantic processing rule represents a transformation rule between the first adjustment semantic information and the second recovery text data;
The receiving end 302 is configured to reply to the first transmission problem with the second recovery text data according to a preset reply rule, and generate a first reply answer corresponding to the first transmission problem, where the reply rule is a rule for calculating the first reply answer corresponding to the first transmission problem according to the specification of the second recovery text data.
In some embodiments, the transmitting end 301 is configured to:
The transmitting end 301 is configured to encode the first answer and the first target answer, respectively, to generate an encoded answer corresponding to the first answer, and an encoded first target answer corresponding to the first target answer;
The sending end performs product processing on the coded reply answer and the coded first target answer to generate a product processing result;
The sending end 301 is configured to obtain a value of a question selection vector corresponding to the first target answer, and perform ratio processing on the product processing result and the value of the question selection vector corresponding to the first target answer, so as to obtain a first semantic similarity between the first answer and the first target answer.
In some embodiments, the transmitting end 301 is configured to:
The transmitting end 301 is configured to obtain a value of a historical first semantic similarity corresponding to a previous adjustment process corresponding to a current adjustment process, obtain a reward value corresponding to the value of the historical first semantic similarity, adjust first adjustment semantic information by using the reward value to obtain adjusted first adjustment semantic information, send the adjusted first adjustment semantic information to the receiving end, so that the receiving end determines a third answer corresponding to the first transmission problem according to the adjusted first adjustment semantic information, and transmit the third answer back to the transmitting end;
The sending end 301 is configured to determine a first semantic similarity between the third answer and the first target answer until a preset first adjustment procedure number is reached, obtain a plurality of first semantic similarities, select a value of the largest first semantic similarity from the values of the plurality of first semantic similarities, and use an adjusted semantic information corresponding to the value of the largest first semantic similarity as second adjustment semantic information.
In some embodiments, the transmitting end 301 is configured to:
The sending end 301 is configured to obtain a preset question selection vector, randomly adjust the question selection vector to obtain an adjusted question selection vector, determine a second transmission question corresponding to the adjusted question selection vector, and a second target answer corresponding to the second transmission question, store the second target answer, and transmit the second transmission question and the first adjustment semantic information to a receiving end, so that the receiving end determines a fourth answer corresponding to the second transmission question according to the first adjustment semantic information;
The sending end 301 is configured to determine a third semantic similarity between the fourth answer and the first target answer until a preset second adjustment procedure number is reached, obtain a plurality of third semantic similarities, select a value of the smallest third semantic similarity from the plurality of third semantic similarities, use an adjusted question selection vector corresponding to the value of the smallest third semantic similarity as a target question selection vector, and obtain a target transmission question corresponding to the target question selection vector;
The transmitting end 301 is configured to obtain a value of a historical first semantic similarity corresponding to a previous adjustment process corresponding to a current adjustment process, obtain a reward value corresponding to the value of the historical first semantic similarity, adjust first adjustment semantic information by using the reward value to obtain adjusted first adjustment semantic information, send the adjusted first adjustment semantic information to the receiving end, so that the receiving end determines a fourth answer corresponding to the first transmission problem according to the adjusted first adjustment semantic information, and transmit the fourth answer back to the transmitting end;
The sending end 301 is configured to determine a first semantic similarity between the fourth answer and the first target answer until a preset third adjustment procedure number is reached, obtain a plurality of first semantic similarities, select a value of the largest first semantic similarity from the values of the plurality of first semantic similarities, and use an adjusted semantic information corresponding to the value of the largest first semantic similarity as second adjustment semantic information.
For convenience of description, the above devices are described as being functionally divided into various modules, respectively. Of course, the functions of each module may be implemented in the same piece or pieces of software and/or hardware when implementing the present application.
The device of the foregoing embodiment is configured to implement the corresponding semantic communication method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiment, which is not described herein.
Based on the same inventive concept, the application also provides an electronic device corresponding to the method of any embodiment, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the semantic communication method of any embodiment when executing the program.
Fig. 4 shows a more specific hardware architecture of an electronic device according to this embodiment, where the device may include: a processor 401, a memory 402, an input/output interface 403, a communication interface 404, and a bus 405. Wherein the processor 401, the memory 402, the input/output interface 403 and the communication interface 404 are in communication connection with each other inside the device via a bus 405.
The processor 401 may be implemented by a general purpose CPU (Central Processing Unit ), a microprocessor, an Application SPECIFIC INTEGRATED Circuit (ASIC), or one or more integrated circuits, etc. for executing related programs to implement the technical solutions provided in the embodiments of the present disclosure.
The Memory 402 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory ), static storage, dynamic storage, etc. Memory 402 may store an operating system and other application programs, and when implementing the solutions provided by the embodiments of the present specification by software or firmware, the relevant program code is stored in memory 402 and invoked for execution by processor 401.
The input/output interface 403 is used to connect with an input/output module to realize information input and output. The input/output module may be configured as a component in a device (not shown) or may be external to the device to provide corresponding functionality. Wherein the input devices may include a keyboard, mouse, touch screen, microphone, various types of sensors, etc., and the output devices may include a display, speaker, vibrator, indicator lights, etc.
The communication interface 404 is used to connect a communication module (not shown in the figure) to enable communication interaction between the present device and other devices. The communication module may implement communication through a wired manner (such as USB, network cable, etc.), or may implement communication through a wireless manner (such as mobile network, WIFI, bluetooth, etc.).
Bus 405 includes a path to transfer information between components of the device (e.g., processor 401, memory 402, input/output interface 403, and communication interface 404).
It should be noted that, although the above device only shows the processor 401, the memory 402, the input/output interface 403, the communication interface 404, and the bus 405, in the implementation, the device may further include other components necessary for realizing normal operation. Furthermore, it will be understood by those skilled in the art that the above-described apparatus may include only the components necessary to implement the embodiments of the present description, and not all the components shown in the drawings.
The electronic device of the foregoing embodiment is configured to implement the corresponding semantic communication method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiment, which is not described herein.
Based on the same inventive concept, the present application also provides a non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the semantic communication method according to any of the embodiments above, corresponding to the method of any of the embodiments above.
The computer readable media of the present embodiments, including both permanent and non-permanent, removable and non-removable media, may be used to implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device.
The storage medium of the foregoing embodiments stores computer instructions for causing the computer to perform the semantic communication method of any of the foregoing embodiments, and has the advantages of the corresponding method embodiments, which are not described herein.
Those of ordinary skill in the art will appreciate that: the discussion of any of the embodiments above is merely exemplary and is not intended to suggest that the scope of the application (including the claims) is limited to these examples; the technical features of the above embodiments or in the different embodiments may also be combined within the idea of the application, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the application as described above, which are not provided in detail for the sake of brevity.
Additionally, well-known power/ground connections to Integrated Circuit (IC) chips and other components may or may not be shown within the provided figures, in order to simplify the illustration and discussion, and so as not to obscure the embodiments of the present application. Furthermore, the devices may be shown in block diagram form in order to avoid obscuring the embodiments of the present application, and also in view of the fact that specifics with respect to implementation of such block diagram devices are highly dependent upon the platform within which the embodiments of the present application are to be implemented (i.e., such specifics should be well within purview of one skilled in the art). Where specific details (e.g., circuits) are set forth in order to describe example embodiments of the application, it should be apparent to one skilled in the art that embodiments of the application can be practiced without, or with variation of, these specific details. Accordingly, the description is to be regarded as illustrative in nature and not as restrictive.
While the application has been described in conjunction with specific embodiments thereof, many alternatives, modifications, and variations of those embodiments will be apparent to those skilled in the art in light of the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.
The present embodiments are intended to embrace all such alternatives, modifications and variances which fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, and the like, which are within the spirit and principles of the embodiments of the application, are intended to be included within the scope of the application.

Claims (10)

1. A semantic communication method, applied to a semantic communication system, the system comprising a transmitting end and a receiving end, the method comprising:
The sending end obtains text data to be transmitted, generates a first transmission question and first adjustment semantic information associated with the text data to be transmitted based on the text data to be transmitted, determines a first target answer corresponding to the first transmission question, stores the first target answer, and transmits the first adjustment semantic information and the first transmission question to the receiving end;
the receiving end receives first adjustment semantic information and a first transmission problem transmitted by the transmitting end, determines a first reply answer corresponding to the first transmission problem based on the first adjustment semantic information, and transmits the first reply answer to the transmitting end;
The sending terminal receives a first answer transmitted by the receiving terminal, determines first semantic similarity between the first answer and the first target answer, and adjusts the first adjustment semantic information according to the first semantic similarity to generate second adjustment semantic information;
The receiving end receives second adjustment semantic information transmitted by the transmitting end, performs semantic processing according to the second adjustment semantic information, and generates first recovery text data, wherein the first recovery text data is identical to the text data to be transmitted.
2. The method of claim 1, wherein the generating a first transmission question and first adjustment semantic information associated with the text data to be transmitted based on the text data to be transmitted comprises:
the sending end inputs the text data to be transmitted into a pre-trained problem generation model, and generates a first transmission problem associated with the text data to be transmitted through the problem generation model;
The sending end extracts initial semantic information from text data to be transmitted according to a preset first semantic processing rule, continuously adjusts the initial semantic information to obtain adjusted initial semantic information, and sends the adjusted initial semantic information to the receiving end in the continuous adjustment process so that the receiving end can determine a second answer corresponding to the first transmission problem according to the adjusted initial semantic information and send the second answer back to the sending end;
The sending end determines second semantic similarity between the second answer and the first target answer until the value of the second semantic similarity is the same as the value of the initial semantic similarity determined randomly, and takes the adjusted initial semantic information corresponding to the second semantic similarity as first adjustment semantic information.
3. The method of claim 1, wherein the determining a first target answer corresponding to the first transmission question comprises:
The sending end inputs the text data to be transmitted and the first transmission problem into a pre-trained answer generation model, and a plurality of answers corresponding to the first transmission problem are generated through the answer generation model;
and the sending end selects a first target answer from the multiple answers according to a preset question selection vector.
4. The method of claim 1, wherein the determining a first answer to a reply to the first transmission question based on the first adjustment semantic information comprises:
The receiving end performs recovery processing according to the first adjustment semantic information by using a preset second semantic processing rule to generate second recovery text data, wherein the second semantic processing rule represents a transformation rule between the first adjustment semantic information and the second recovery text data;
And the receiving end replies the first transmission problem by utilizing the second recovery text data according to a preset reply rule, and generates a first reply answer corresponding to the first transmission problem, wherein the reply rule is a rule for calculating the first reply answer corresponding to the first transmission problem according to the specification of the second recovery text data.
5. The method of claim 1, wherein the determining a first semantic similarity between the first answer to answer and the first target answer comprises:
The sending end encodes the first answer and the first target answer respectively, and generates an encoded answer corresponding to the first answer and an encoded first target answer corresponding to the first target answer;
The sending end performs product processing on the coded reply answer and the coded first target answer to generate a product processing result;
And the sending end acquires the value of the question selection vector corresponding to the first target answer, and carries out ratio processing on the product processing result and the value of the question selection vector corresponding to the first target answer to obtain the first semantic similarity between the first answer and the first target answer.
6. The method of claim 1, wherein adjusting the first adjustment semantic information according to the first semantic similarity generates second adjustment semantic information, comprising:
The sending end obtains a historical first semantic similarity value corresponding to a previous adjusting process corresponding to a current adjusting process, obtains a reward value corresponding to the historical first semantic similarity value, adjusts first adjusting semantic information by using the reward value to obtain adjusted first adjusting semantic information, sends the adjusted first adjusting semantic information to the receiving end, so that the receiving end determines a third answer corresponding to the first transmission problem according to the adjusted first adjusting semantic information, and returns the third answer to the sending end;
The sending end determines the first semantic similarity between the third answer and the first target answer until the preset first adjustment process times are reached, a plurality of first semantic similarities are obtained, the value of the largest first semantic similarity is selected from the values of the plurality of first semantic similarities, and the adjusted semantic information corresponding to the value of the largest first semantic similarity is used as second adjustment semantic information.
7. The method of claim 1, wherein adjusting the first adjustment semantic information according to the first semantic similarity generates second adjustment semantic information, comprising:
The sending end obtains a preset question selection vector, randomly adjusts the question selection vector to obtain an adjusted question selection vector, determines a second transmission question corresponding to the adjusted question selection vector, and a second target answer corresponding to the second transmission question, stores the second target answer, and transmits the second transmission question and the first adjustment semantic information to the receiving end so that the receiving end can determine a fourth answer corresponding to the second transmission question according to the first adjustment semantic information;
The sending end determines third semantic similarity between the fourth answer and the first target answer until the preset second adjustment process times are reached, a plurality of third semantic similarities are obtained, the value of the smallest third semantic similarity is selected from the plurality of third semantic similarities, an adjusted question selection vector corresponding to the value of the smallest third semantic similarity is used as a target question selection vector, and a target transmission question corresponding to the target question selection vector is obtained;
The sending end obtains a historical first semantic similarity value corresponding to a previous adjusting process corresponding to a current adjusting process, obtains a rewarding value corresponding to the historical first semantic similarity value, adjusts first adjusting semantic information by utilizing the rewarding value to obtain adjusted first adjusting semantic information, sends the adjusted first adjusting semantic information to the receiving end, so that the receiving end determines a fourth answer corresponding to the first transmission problem according to the adjusted first adjusting semantic information, and returns the fourth answer to the sending end;
The sending end determines the first semantic similarity between the fourth answer and the first target answer until the preset third adjustment process times are reached, a plurality of first semantic similarities are obtained, the value of the largest first semantic similarity is selected from the values of the plurality of first semantic similarities, and the adjusted semantic information corresponding to the value of the largest first semantic similarity is used as second adjustment semantic information.
8. A semantic communication apparatus, the apparatus being provided in a semantic communication system, the system comprising a transmitting end and a receiving end, the apparatus comprising:
The sending end is configured to acquire text data to be transmitted, generate a first transmission problem and first adjustment semantic information associated with the text data to be transmitted based on the text data to be transmitted, determine a first target answer corresponding to the first transmission problem, store the first target answer, and transmit the first adjustment semantic information and the first transmission problem to the receiving end;
The receiving terminal is configured to receive first adjustment semantic information and a first transmission question transmitted by the transmitting terminal, determine a first answer corresponding to the first transmission question based on the first adjustment semantic information, and transmit the first answer to the transmitting terminal;
The sending end is configured to receive a first answer transmitted by the receiving end, determine a first semantic similarity between the first answer and the first target answer, and adjust the first adjustment semantic information according to the first semantic similarity to generate second adjustment semantic information;
The receiving end is configured to receive the second adjustment semantic information transmitted by the transmitting end, perform semantic processing according to the second adjustment semantic information, and generate first recovery text data, wherein the first recovery text data is the same as the text data to be transmitted.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the method of any one of claims 1 to 7 when the program is executed by the processor.
10. A non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the method of any one of claims 1 to 7.
CN202410058360.0A 2024-01-15 2024-01-15 Semantic communication method, semantic communication device, electronic equipment and storage medium Pending CN118138192A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410058360.0A CN118138192A (en) 2024-01-15 2024-01-15 Semantic communication method, semantic communication device, electronic equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410058360.0A CN118138192A (en) 2024-01-15 2024-01-15 Semantic communication method, semantic communication device, electronic equipment and storage medium

Publications (1)

Publication Number Publication Date
CN118138192A true CN118138192A (en) 2024-06-04

Family

ID=91233531

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410058360.0A Pending CN118138192A (en) 2024-01-15 2024-01-15 Semantic communication method, semantic communication device, electronic equipment and storage medium

Country Status (1)

Country Link
CN (1) CN118138192A (en)

Similar Documents

Publication Publication Date Title
CN111309883A (en) Man-machine conversation method based on artificial intelligence, model training method and device
CN113961967A (en) Method and device for jointly training natural language processing model based on privacy protection
US12106555B2 (en) Method and device for retrieving image
CN117910518A (en) Method and system for analyzing generated data
CN113723115B (en) Open domain question-answer prediction method based on pre-training model and related equipment
CN118113815A (en) Content searching method, related device and medium
CN116306704B (en) Chapter-level text machine translation method, system, equipment and medium
CN118138192A (en) Semantic communication method, semantic communication device, electronic equipment and storage medium
CN115796156A (en) Text error correction method, device, equipment and medium
CN116030798A (en) Pre-training method of audio encoder, audio detection method and device
CN113111642B (en) Natural language recognition model generation and natural language processing method and equipment
WO2018135281A1 (en) Data compression device, data compression program, and data compression method and program
CN115374766A (en) Text punctuation recovery method and related equipment
CN110929767B (en) Font processing method, system, device and medium
CN114417251A (en) Retrieval method, device, equipment and storage medium based on hash code
US20230042275A1 (en) Network quantization method and network quantization device
CN117974833B (en) Image generation method, apparatus, device, readable storage medium, and program product
CN117271803B (en) Training method, device, equipment and storage medium for knowledge graph completion model
CN110727442B (en) Data storage optimization method and system for embedded platform
CN118260155A (en) Micro-service monitoring method based on service center and related equipment
CN114186043B (en) Pre-training method, device, equipment and storage medium
CN117853316A (en) Character portrait generation method, system, electronic equipment and storage medium
CN118842724A (en) Digital twin power network communication method, device, electronic equipment and storage medium
CN117932475A (en) Matching method of grafting data source and related equipment
US20210019625A1 (en) Electronic device and control method thereof

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination