User-defined function identification method and device based on knowledge enhancement , and medium

Abstract

The invention discloses a user-defined function recognition method and device based on knowledge enhancement, and a medium. The method comprises the following steps: carrying out the preprocessing and division processing of a disassembly text of a user-defined function, and obtaining a training data set; performing word embedding on the operation codes to generate word vectors; inputting the word vectors into a Transform model, and obtaining a Transform model which is subjected to initial pre-training; performing mask processing on the word vectors with a preset number threshold value; and inputting all word vectors into the Transform model subjected to initial pre-training, outputting a coding result of the word vectors, taking the coding result as coding feature information of the disassembled text, converting the statistical features into a three-channel image to obtain statistical feature information of the disassembled text, splicing the statistical feature information and the coding feature information, inputting into a full-connection neural network, and obtaining a user-defined function classification model. According to the embodiment of the invention, the user-defined function can be accurately identified.

Description

technical field [0001] The present invention relates to the field of computer technology, in particular to a user-defined function recognition method, device and medium based on knowledge enhancement. Background technique [0002] At present, the vast majority of software is released in the form of executable binary code rather than source code, which can be used by mobile phones and computers. There is a huge difference between the binary code and the source code, not only in the lack of readability of the binary code, but also in the fact that the reverse analysis of the binary code can dig out the loopholes that are difficult to analyze or cannot be found in the source code. However, due to the existence of binary code loopholes, the software is extremely vulnerable to malicious attacks, which seriously damages the quality of the software, especially for non-open source software that only discloses binary codes, there may be more loopholes that have not been discovered or...

AI Technical Summary

Problems solved by technology

However, due to the existence of binary code loopholes, the software is extremely vulnerable to malicious attacks, which seriously damages the quality of the software, especially for non-open source software that only discloses binary codes, there may be more loopholes that have not been discovered or fixed. higher security risk

Images