Research On Fuzzy Testing Technology Based On Deep Learning

With the continuous progress and development of global Internet technology,network applications have gradually integrated into everyone's daily life,but the ability of developers is different,which will lead to the existence of program bugs or security loopholes in network applications,which will affect the people's online security environment.Therefore,it has become a top priority to effectively discover the security vulnerabilities in web applications,and on this basis,assist developers to repair them in time.At present,the traditional fuzzy testing method of web application vulnerability has the problems of poor effectiveness,high randomness and low aggressiveness of test cases,so it has very important theoretical significance and practical value to generate test cases purposefully,improve the aggressiveness of test cases and enhance the efficiency of fuzzy testing of web vulnerability.In this paper,the current fuzzy test technology is deeply analyzed and studied,and some shortcomings are found.Aiming at the shortcomings,this paper proposes a method to generate web application test cases based on improved generative adversarial networks.In this method,the attack of test cases is enhanced by using the generative adversarial networks countermeasure training.The TCSA is added and the leakgan model is improved to make it suitable for web vulnerability fuzzy test case generation.On the basis of the previous improvements,a more efficient fuzzy tester is realized.Finally,a simulation environment is designed to simulate the normal fuzzy test environment.A series of experiments show that the method of generating test cases based on improved generative adversarial networks algorithm proposed in this paper has strong practicability in fuzzy test.The main contents of this paper are as follows:Firstly,the method of generating test cases based on improved generative adversarial networks is proposed.Based on the research of generative adversarial networks algorithm,the leakgan model is optimized and improved according to the previous experience,and the TCSA is added to generate test cases better.Through the continuous training of the improved leakgan model,including the continuous training of generator and discriminator,finally,the improved model generates better array test cases.Compared with genetic algorithm and leakgan model,the test cases generated by this method are verified to have high integrity,and then simulation experiments are carried out to verify the effectiveness of the generated test cases.Secondly,a more efficient fuzzy tester is designed and developed,and each module is introduced and implemented.The fuzzy tester can use the improved leakgan model to generate web vulnerability fuzzy test cases and mine web application vulnerabilities.Thirdly,the emulation environment similar to the real fuzzy test environment is built to verify the effectiveness and functionality of the proposed method.During the experiment,the experiment was recorded in detail and the results were analyzed.According to the final results of the experiment,we find that the fuzzy tester based on the improved generative adversarial networks algorithm can increase the effectiveness of test cases to a certain extent.In this experiment,we can effectively mine XSS cross site script attack vulnerability by generating test cases.The experimental results of the method and the system we designed show that the test cases generated based on the improved generative adversarial networks algorithm can effectively increase the availability of test cases,and can improve the efficiency of fuzzy testing to a certain environment;the fuzzy tester based on the improved generative adversarial networks algorithm can effectively mine the loopholes in the actual web application by using the method in this paper,and the experimental results are satisfactory It is proved that it is feasible and practical.