124results about How to "Increase randomness" patented technology

The invention provides a privacy data analysis method and system based on collaborative learning and dynamic encryption, and relates to the technical field of privacy, and the method comprises the steps: receiving original data of a plurality of participants, the original data of each participant being in an initial encryption state; performing secondary encryption on the original data of each participant by using the dynamic encryption key to obtain secondary encryption data; inputting the secondary encrypted data into a preset collaborative learning model, and decrypting the collaborative learning model by using a security computing protocol under homomorphic encryption in combination with a dynamic encryption key of each participant to obtain a privacy protection analysis model which can be used by each participant; when data analysis is carried out, the privacy protection analysis model is used for processing input data to be analyzed, and correctness of the analysis process is ensured through a verifiable computing technology; and based on a predefined access control strategy, only returning an analysis result that the participant has access to the participant to the corresponding participant.

The invention discloses an unmanned aerial vehicle path planning method based on global search and meta-heuristic optimization, and the method comprises the steps: firstly building an unmanned aerial vehicle path planning model under complex constraints, determining a fitness function and a constraint condition thereof, setting algorithm initialization parameters, carrying out the chaotic mapping of an initialization population position, and carrying out the optimization of the algorithm; and performing loop iteration solution by using a meta-heuristic optimization algorithm, outputting a global optimal path, finally realizing the global optimal path on the unmanned aerial vehicle, and calculating a flight distance corresponding to the optimal path to complete unmanned aerial vehicle path planning. According to the method, a meta-heuristic algorithm, namely a crown porcupine algorithm, is combined with tabu search, the global search capability is enhanced, and meanwhile, population initialization is optimized in combination with chaotic mapping, so that the randomness and unpredictability are increased, the algorithm can converge more quickly, and the robustness of the algorithm is improved. The method has obvious advantages in solving the problem of path planning with complex environment and complex constraint, and solves the problem of path planning of the unmanned aerial vehicle in the three-dimensional complex mountain environment.

The invention discloses an environmental monitoring system based on an optical sensor, and relates to the technical field of environmental monitoring, and the system comprises the following components: a quantum entanglement module, an optical sensor integration module, a quantum measurement and data processing module, and a data transmission and sharing module. By integrating the quantum entanglement module, the optical sensor integration module and the quantum measurement and data processing module, high-precision and real-time monitoring of environmental parameters is realized, the quantum entanglement technology provides measurement precision and anti-interference capability superior to those of a traditional optical sensor for the system, the accuracy and reliability of data are ensured, and meanwhile, the quantum entanglement technology can be used for monitoring the environmental parameters in real time. Due to the design of high sensitivity and wide dynamic range of the optical sensor, the system can stably work under various complex environment conditions, and the efficiency and accuracy of environment monitoring are effectively improved.

The invention discloses a file transmission protection method, system and device based on multiple encryption algorithms, and relates to the technical field of information, and the method comprises the steps: selecting a target file needing to be transmitted, carrying out the preprocessing, and generating file fragments; analyzing the sensitivity of each file fragment, and performing hierarchical encryption processing to obtain an encrypted fragment set; selecting a steganography carrier file, and embedding the encrypted fragments into the carrier file to generate a steganography index table; generating a dynamic session key based on the confusion key pool, encrypting the steganographic file transmission layer and embedding a dynamic key generation parameter; the receiving end generates a session key according to the transmission message header, decrypts the transmission layer data, extracts the steganographic file and decrypts the steganographic file; recombining the decrypted and recovered fragmented data into a complete file according to an identifier sequence; by dynamically matching the steganography carrier file, the efficient embedding of the fragmented data and the generation of the steganography index table are realized, and the concealment and the adaptability in the transmission process are ensured.

The invention relates to a data calibration method suitable for an atmospheric pollutant sensor deployed in an embedded terminal. The method comprises the following steps: acquiring original response data of different sensors with different concentrations in different atmospheric environments after temperature compensation, and corresponding temperature, humidity, air pressure data and sampling time; collecting real concentration values of detection objects corresponding to different sensors at corresponding moments given by a national control station or a provincial control station; preprocessing the obtained data to obtain a final data set; a random forest algorithm is used as a teacher model, a gradient elevator algorithm is used as a student model, and hyper-parameters of the gradient elevator algorithm and the random forest algorithm are optimized by using a hedera helix algorithm; substituting a hyper-parameter optimal solution found by optimization into a corresponding algorithm, and retraining the teacher model and the student model by using hyper-parameter combinations; and deploying, updating and displaying the model in the embedded equipment. And the model size is reduced to realize the deployment of the embedded end on the premise of ensuring the accuracy.

The invention discloses a dynamic S-box generation method and test system based on a spatiotemporal chaotic system, and the method comprises the steps: constructing an iteration matrix through employing a preset m sequence, and determining a coupling object index and a disturbance symbol of each grid in an iteration process; wherein each element of the iteration matrix represents a grid; performing dynamic disturbance on the initial chaotic sequence according to the coupling object index, the disturbance symbol, the index offset and the grid coordinate to obtain a dynamic chaotic sequence; initializing the two-dimensional coupling image lattice by using a dynamic chaos sequence, and carrying out iteration to obtain an iterated coupling image lattice; removing the previous N iteration results of the coupled image lattice after iteration; and performing numerical conversion on the chaotic sequence value obtained by the (N + 1) th iteration to construct an S box. According to the S box design and test method, the safety and the attack resistance of the S box are improved, and an efficient and reliable research tool can be provided for cryptography researchers.

The invention discloses an AGV obstacle avoidance path planning method in an unmanned warehousing system, and the method comprises the steps: employing nested Logistic mapping and Skew Tent mapping to generate a chaotic initialization population to replace a population generation mode of a conventional differential evolution algorithm, carrying out the Levy flight operation after the selection operation, and carrying out the global path search on an environment map, thereby achieving the optimal path planning. And obtaining an AGV global feasible path avoiding obstacles in the environment map, and when the AGV has an unknown obstacle on the feasible path, re-planning the local path by adopting a dynamic window method so as to enable the AGV to realize an obstacle avoidance process in an interval where the unknown obstacle is located. The improved differential evolution algorithm enhances the global search capability through chaos initialization and a Levy flight strategy, and the AGV can quickly respond to an unknown obstacle through a dynamic window rule. The adaptability and the accuracy of path planning are improved, and the obstacle avoidance capability and the safety of the AGV in a complex environment are enhanced at the same time.

The invention relates to a multi-unmanned aerial vehicle cooperative confrontation decision-making method and system. The method comprises the following steps: constructing a multi-unmanned aerial vehicle cooperative air combat confrontation decision-making environment by establishing a multi-unmanned aerial vehicle air combat confrontation motion model and an air combat situation evaluation model; establishing a distributed partially observable Markov decision process model of the multi-unmanned aerial vehicle cooperative confrontation decision problem according to the action space and the local observation and state of each unmanned aerial vehicle in the confrontation decision environment; designing a multi-machine collaborative adversarial reward function and an HASAC algorithm network space; and training and generating a multi-unmanned-aerial-vehicle cooperative confrontation strategy model based on interaction between a network space and a multi-unmanned-aerial-vehicle cooperative confrontation decision-making environment of an HASAC algorithm. According to the method, for the multi-aircraft cooperative air combat confrontation problem, the specific global state is designed for the multi-aircraft cooperative confrontation decision-making unmanned aerial vehicle, and compared with direct splicing of observation vectors of all unmanned aerial vehicles, the global state dimension is reduced, and the training efficiency is improved.

The invention discloses a network edge calculation method and a communication device, and relates to the technical field of network communication, and the method comprises the following calculation steps: inputting original data into a calculation processing module, and outputting processed data, encrypted data, a hash value of the encrypted data, an authentication code of a message and a verification result. And transmitting the encrypted data, the hash value of the encrypted data, the authentication code of the message and the verification result to a cloud server. The method has the advantages that the received data is changed firstly, so that the encryption efficiency and security of the data are improved, the data are encrypted in an encryption mode, the confidentiality of the data is guaranteed, the integrity verification of the data is carried out through the hash function and the message authentication code, and the security of the data is improved. According to the data transmission method and device, the data are ensured not to be tampered in the transmission process, all-directional protection of the data is achieved, meanwhile, the received data are verified in a verification mode so as to ensure the authenticity of the data, and dual guarantee of data transmission and verification is achieved.

The invention discloses a mixed sampling network traffic sample balancing method, which comprises the following steps of: firstly, carrying out data preprocessing on network traffic in a CIC IDS-2017 data set, and dividing the network traffic into a training set and a test set; secondly, the training set is subjected to SMOTE oversampling and ENN undersampling processing, Gaussian noise processing is introduced after SMOTE oversampling, and a mixed sampling sample is obtained; and finally, carrying out CNN training on the mixed sampling sample and verifying a training result through a test set to obtain a classification result. According to the method, Gaussian noise is added after SMOTE processing, so that the complexity of data is increased, the generated sample can better reflect various characteristics in real network traffic, and the risk of overfitting is remarkably reduced. According to the method, most of important information is reserved in the ENN, and unnecessary redundant data is reduced, so that the problem of information loss is avoided, and the adaptability of the model to complex network traffic data is improved. And the function of balancing the network flow data is realized.

The invention discloses a multi-dimensional load spectrum acceleration compilation method for a vehicle frame based on a WLTC working condition, and the method comprises the following steps: setting rigid connection of the vehicle frame, and building a whole vehicle dynamics model; creating a WLTC event working condition, carrying out whole vehicle dynamics simulation, and extracting a load at a connection point; performing finite element analysis by using an inertial release method; carrying out fatigue life analysis on the frame, and extracting fatigue hot spot numbers; reconstructing an analysis process, and extracting stress time sequence data; carrying out pseudo damage analysis on the stress time spectrum to intercept an original load spectrum; carrying out smooth connection processing on the intercepted acceleration spectrum fragments; and performing power spectrum density analysis on the acceleration spectrum and the original spectrum, screening the acceleration spectrum according to a result, and finally obtaining the multi-dimensional load acceleration spectrum of the vehicle frame based on the WLTC working condition, thereby shortening the bench test time and reducing the development cost of the vehicle frame.

The embodiment of the invention provides an automatic driving scene data processing method and device and a storage medium. At least one traffic track on the road section can be identified according to dynamic traffic data in the roadside sensing data; carrying out key frame extraction on the at least one traffic track to obtain at least one group of key frames; performing decision planning according to the at least one group of key frames and a preset constraint condition to obtain at least one driving track; and determining automatic driving scene data according to static road network information in the roadside sensing data and the at least one driving track. Therefore, the automatic driving scene data in the rich scene can be obtained without actual measurement of the automatic driving vehicle on the road. Besides, variability and randomness of the automatic driving scene data can be increased by utilizing settable characteristics of constraint conditions, so that driving tracks in various different scenes are derived, and the coverage rate of the automatic driving scene data to different automatic driving scenes is increased.

The invention relates to the technical field of power system optimization, and discloses a multi-objective particle swarm optimization method and system based on multi-strategy improvement, and the method comprises the steps: solving a multi-objective optimization model of a power supply of a generator state monitoring device through employing a multi-strategy improved multi-objective particle swarm optimization algorithm; three improvement strategies of self-adaptive adjustment of inertia weight, coexistence of a decomposition algorithm and Pareto dominance and introduction of variation factors are combined; defects of a traditional multi-target particle swarm algorithm are overcome, and better distribution of Pareto leading edges is realized, so that an optimal Pareto optimal solution set is obtained. Compared with the prior art, the improved multi-objective particle swarm optimization method provided by the invention solves the problems that a local non-dominated solution is converged too early when a traditional multi-objective particle swarm optimization algorithm is used for solving a multi-objective optimization problem, and Pareto leading edge suboptimal distribution is caused by an improper external archive updating strategy; therefore, the operation efficiency and reliability of the power supply system of the device are improved.

The invention belongs to the field of wireless communication, and discloses a power wireless local area network information security transmission method and a related device, and the method comprises the steps: obtaining a current physical layer leader sequence and a physical layer header, and carrying out the processing of to-be-transmitted data on the basis of the current physical layer leader sequence and the physical layer header, thereby achieving the dynamic operation of the physical layer leader sequence and the physical layer header. Meanwhile, when data to be transmitted are modulated by adopting an orthogonal frequency division multiplexing technology, a positive value random quantity is introduced to modify a cyclic prefix selection area and length and modify subcarrier parameters, so that the randomness and unpredictability of signal transmission are further improved, and the hiding of physical layer waveform parameters is realized. By dynamically switching the waveform factors of the physical layer, an attacker is difficult to capture and identify the characteristic information of the wireless signal due to the dynamic change, so that the anti-reconnaissance and anti-interception capabilities of the system are greatly improved. A parameter dynamic change rule can be self-formulated for the power wireless local area network, so that the powerful supervision capability of the power industry on the wireless local area network technology can be realized.

The invention provides a route planning method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a state parameter of a vehicle and a target position parameter of a destination of the vehicle, and determining a reward function of an estimation neural network and an action space of the vehicle according to the state parameter and the target position parameter; inputting the state parameters and the action space into an estimation neural network to obtain a plurality of estimation values corresponding to the plurality of space directions; obtaining the number of repeated selection times of each space direction in the plurality of space directions in the estimation neural network, and performing parameter adjustment on the plurality of estimation values according to the number of repeated selection times to obtain a plurality of target values corresponding to the plurality of space directions; and if there is a target value greater than the threshold value in the plurality of target values, determining a target space direction corresponding to the target value greater than the threshold value, and performing route planning according to the target space direction and the speed parameter. The method can reduce the times of repeatedly adopting actions, and prevents the training of the estimated neural network from tending to a local optimal solution.

The invention provides a data protection method and device for a fingerprint browser and a medium, and the method comprises the steps: inputting device fingerprint data, and constructing a device fingerprint data set; calculating an entropy value of each feature according to the high-frequency feature set and the low-frequency feature set, and constructing an entropy value set of each feature; designing a low-frequency feature disturbance formula for the low-frequency feature set to obtain a disturbed low-frequency feature value and a disturbed low-frequency feature set; constructing a generative adversarial network according to the high-frequency feature set and the disturbed low-frequency feature set; and collecting dynamic feedback information of the tracking system in real time, and finally outputting an updated generative adversarial network and a verified confusion feature set. According to the method, the contradiction between privacy protection and function compatibility in the prior art is solved, and the dynamic adaptation capability of the system is remarkably improved.

The invention discloses an image encryption method and device based on a coupled image lattice and a cellular automaton, and relates to the field of image processing, and the method comprises the steps: constructing a coupled image lattice chaotic system of an improved coupled structure based on an elementary cellular automaton and two optimized one-dimensional mappings; performing correlation calculation on the plaintext image and the external key to generate an internal key; inputting the internal key into the coupling mapping lattice chaotic system of the improved coupling structure for iteration and expansion to generate a chaotic sequence matrix; selecting and recombining the chaotic sequence matrix to generate an initial cellular matrix and a key matrix; scrambling the plaintext image according to the initial cellular matrix by adopting a scrambling algorithm based on a two-dimensional cellular automaton to obtain a scrambled image; and diffusing the scrambled image according to the key matrix by adopting a diffusion algorithm based on a class cellular automaton to obtain a ciphertext image. The method and the device can simplify operation and improve encryption efficiency and safety performance.

The invention discloses a jitter increasing circuit and a true random number generator. The jitter increasing circuit comprises a current random fluctuation module and an adjustable noise module; the current output end of the current random fluctuation module is connected with the drain electrode and the grid electrode of the field effect transistor M0, and the current random fluctuation module can generate random current; the grid electrode of the field effect transistor M0 is connected with the grid electrode of the field effect transistor M1; the drain electrode of the field effect transistor M1 is grounded through a capacitor CL; the adjustable noise module is arranged beside the capacitor CL, and noise generated by the adjustable noise module can be coupled and superposed to the capacitor CL; the drain electrode of the field effect transistor M1 is connected with the high-frequency clock circuit through the phase inverter; the grid electrode of the field effect transistor M3 is connected with the clock source circuit through the phase inverter. According to the invention, a function of sampling clock random jitter increase can be realized on a common chip, noise in two dimensions of voltage and current can be superposed, the randomness of sampling clock jitter is greatly improved, and the randomness of TRNG output is further improved.

The invention discloses a multi-view MIMO radar correlated imaging radiation field design method based on image entropy, which considers the fluctuation of the scattering intensity of each resolution unit of a target in an imaging scene under a multi-view condition, carries out statistical modeling on the scattering intensity of each resolution unit of the target, and starts from the whole process of transmitting, receiving and processing a signal. By constraining the condition number of a radiation field Gram matrix and the image entropy of imaging and through continuous iterative optimization, an array type of an MIMO radar transmitting waveform and a transmitting array which have a good suppression effect on target grid energy dissipation and are stable in imaging is found, and the correlation between a radiation field reference matrix and signal echoes is improved. According to the method, radar correlated imaging can be carried out on the target with fluctuating scattering intensity of each resolution unit by using the optimized emission signal waveform and the array of the emission array, the randomness of a radiation field is enhanced, the energy dissipation degree of a target grid is reduced, the imaging error is reduced, and a better radar imaging result can be obtained.

The invention provides a method and device for displaying a bullet screen special effect, electronic equipment, a computer readable medium and a computer program product. The method comprises the following steps: acquiring bullet screen special effect information corresponding to a played target video; obtaining target bullet screen data, wherein the target bullet screen data comprises a bullet screen which needs to render a special effect; based on the bullet screen special effect information, bullet screen element special effects corresponding to target bullet screen data are displayed, the bullet screen element special effects achieve the effect that a bullet screen forms special effect element backflow with a preset suction center point in a target video image as a center axis, and the positions of the special effect elements are dynamically adjusted along with the change of the backflow progress. According to the invention, for the video played at the WEB end, the bullet screen element special effect is displayed based on the predetermined trigger display logic, the effect that the bullet screen is displayed in a specific scene and element backflow is formed by taking the preset suction center point in the video picture as the center axis is realized, the bullet screen display effect is enriched, and the user experience is improved.

The invention relates to the technical field of workshop production scheduling, in particular to a distributed hybrid flow workshop scheduling method oriented to rocket tank production, and aims to construct a distributed hybrid flow workshop scheduling model including an assembly process by taking the maximum completion time of a minimized product as an objective function. According to the method, a multi-layer coding genetic algorithm framework is designed, and an elite selection strategy and a neighborhood search operator are introduced into the framework. The elite selection strategy ensures that high-quality individuals are reserved in the iteration process, and the convergence speed of the algorithm is increased; a network diagram method is adopted to identify key paths, neighborhood search is carried out on key workpieces, the scheduling bottleneck is optimized, the quality of solutions is improved, and local optimum is avoided. The method not only can greatly improve the scheduling efficiency of the rocket tank, but also has good applicability and universality, and can be widely applied to other types of distributed hybrid flow shop scheduling environments.

The invention discloses an interference communication integrated waveform generation method based on spectrum dispersion-multi-system phase shift keying. The problems of difficult communication information demodulation and high hardware implementation complexity in interference communication integration research are mainly solved. According to the scheme of the invention, the SMSP interference signal is formed by periodically expanding the chirp signal on the time domain, the communication information is modulated on the phase of the chirp signal, and the plurality of chirp signals carrying the communication information can cause dense false target interference at random positions on the radar, so that the interference and communication integrated signal is generated. At a communication receiving end, a first linear frequency modulation signal is adopted to perform carrier synchronization, so that the synchronization effect can be improved by using the high time width-bandwidth product characteristic of the linear frequency modulation signal, and communication information can be demodulated directly by using a coherent demodulation mode; at a radar receiving end, the signal can generate a radar interference effect with dense false target characteristics.

The invention relates to the technical field of coin pushing machines, and discloses a novel pushing disc for a coin pushing machine, which comprises a coin pushing assembly, the coin pushing assembly comprises a fixing frame, wherein the rear surface of the fixing frame is slidably connected with a pushing disc; the baffle is fixedly connected to the side face of the push disc; a sliding assembly is arranged on the front surface of the fixing frame. The sliding assembly comprises a motor which is fixedly connected to the front surface of the fixing frame; the motor gear is meshed with the rack to drive the push disc to reciprocate by a fixed distance. When the coin pushing machine has a big award, the pushing disc moves through the first photoelectric sensor to the third photoelectric sensor, the first photoelectric sensor and the third photoelectric sensor send signals to the coin pushing machine control system, the rack drives the pushing disc to reciprocate by the maximum distance, the pushing disc is reset after reciprocating sliding once, and the motor is driven by the positive and negative rotation driving circuit for control. Working is conducted by combining gear and rack meshing transmission, movement of different distances can be generated, and the interestingness and randomness of games are improved.

The invention relates to a security processing data solidification method and system based on mass data, and belongs to the technical field of data storage processing. The method comprises the following steps: acquiring to-be-stored data, and performing encryption according to the to-be-stored data through an improved AES-ECC hybrid encryption algorithm to obtain to-be-stored encrypted data; constructing a block chain network, and storing the data to block chain nodes through a PBFT consensus algorithm according to the to-be-stored encrypted data; the stored encrypted data is backed up to a plurality of block chain nodes, and curing processing is performed through the data snapshot, so that strong encryption protection in the data storage and transmission process is realized, data leakage and unauthorized access are effectively prevented, data integrity is ensured, and data traceability is realized; meanwhile, data loss prevention is achieved, recovery can be carried out when a system fault or data damage occurs, the block chain network can be expanded according to needs, and the storage and processing requirements of a large amount of data are met.

The invention discloses a secure transmission method and device, equipment, a medium and a product, and the method comprises the steps: obtaining an access request of a client and a current randomized tag pre-distributed to a web application in a server, and triggering the updating of the current randomized tag through a network attack association feature; scheduling a pre-trained target detection model based on the access request, and determining a security detection result of the access request; if the security detection result is secure, sending the access request to a server, and receiving a response result fed back by the server; and performing behavior analysis on the response result according to the current randomized tag, determining security feedback content and feeding back the security feedback content to the client. By dynamically updating the current randomized tag, the randomness of the defense form is enhanced, the method does not depend on the attack mode of an attacker, novel attacks or attack variants can be dealt with in real time, double attack detection of security detection and behavior analysis is carried out on the access request, and the detection capability and defense capability of the attacks are improved.

The application discloses a kind of beam bridge girder static load deflection prediction methods based on improved AVOA-GRNN network, comprising:1, establish vehicle bridge finite element model and add vehicle bridge uncertainty simulation bridge dynamic static load test, collect dynamic static load test data;2, the above data is selected using the characteristic of pearson linear correlation coefficient;3, using dynamic static load experimental data training GRNN neural network model, and construct the parameter of improved African vulture algorithm optimization GRNN, obtain the corresponding smooth factor when average absolute error is minimum;4, the dynamic load test data of prediction set is substituted into improved AVOA-GRNN network model and carries out static load deflection prediction.The application can use beam bridge girder dynamic load test data to predict corresponding static load deflection, so as to effectively solve the problems, such as high cost of predicting beam bridge static load test, great influence on traffic and easily causing damage to bridge structure.

The present application relates to the technical field of food raw material sampling detection, and in particular to a food raw material primary processing sampling detection device, which comprises a conveying bottom plate, the top of the conveying bottom plate is fixed with a horizontal sliding rail, the horizontal sliding rail is connected with a moving plate through an electric sliding block, the top of the moving plate is connected with a containing box for containing raw materials through a T-shaped plate, the T-shaped plate is set upside down, the horizontal section of the T-shaped plate is symmetrically provided with sampling parts in front and back, and the top of the moving plate is commonly provided with reciprocating pushing parts on the left and right sides. The sampling pipe is used to sample from the bottom to the top through the sampling hole preset at the bottom of the containing box, compared with the direct sampling mode of the container from the top of the containing box, the damage to the peas can be reduced, so that the appearance of the processed peas is ensured, and through the pushing arc surface arranged on the side of the sealing plate, the damage to the peas caused by the sealing plate in the process of sealing the through hole can be avoided, and the appearance of the processed peas is further ensured.

The invention discloses a high-confidentiality anti-cheating raw material quality management and control system and a control method. The system comprises a quality system and a supervision system which are matched with each other, wherein the quality system systematically records all operation links in the whole process, and the supervision system monitors all operation links in the whole process. According to the high-confidentiality anti-cheating raw material quality management and control system and control method provided by the invention, the man-made operability is effectively reduced, the falsification behavior is attacked, the risk prevention and control capability of the quality inspection process of the original combustion auxiliary materials entering the factory and the controllability of the inspection process are improved, the real quality of the materials is restored, and the quality of the raw materials is improved. And accurate data support is provided for company production materials.

The application discloses a rumor detection framework based on adaptive data enhancement and adversarial training, and aims to overcome the problems that the prior art only considers structure and ignores features when facing data noise problems, and does not consider the influence of adversarial attacks on rumor detection problems. The steps of the method are as follows: 1. processing the data set; 2. using an adaptive data enhancement method to weaken the influence of noise; 3. using a graph convolutional neural network to generate a graph representation; 4. using an adversarial generation module to generate an adversarial representation to improve the robustness of the model; and 5. constructing a joint loss function to optimize the rumor detection task.