A knowledge graph construction and query recommendation system in the field of radio signal attack and defense

Abstract

A knowledge map construction and query recommendation system in the field of radio signal attack and defense, including: an ontology map design module, a data collection module, a knowledge map construction module, and a query and recommendation module. Algorithms such as defense detection are associated to form a knowledge map, which is convenient for knowledge storage and query. The formed knowledge map includes models, radio signals, algorithms and some derived models and signals, which is beneficial to mining the association in the field of radio signal modulation type identification. The similarity between models and the similarity between signals is conducive to the recommendation of offensive and defensive strategies based on knowledge graphs. Through the recommendation of offensive and defensive strategies, the recommended attack and defense strategies are given for unknown radio signals or classification models, which is beneficial to improve the quality of radio signals. AI Safety in the Domain of Modulation Type Classification.

Description

technical field [0001] The invention relates to the field of radio signal attack and defense, in particular to a knowledge map construction and query recommendation system in the field of radio signal attack and defense. Background technique [0002] In daily life, the use of radio signals for information transmission is still the mainstream method of modern information transmission. After the radio signal is received, it needs to identify its specific modulation type, and then perform corresponding demodulation and extraction according to the modulation type. For the information carried by the signal, people usually need to spend a lot of time and energy to collect data samples of known categories, and use classification algorithms to train these data samples to form a clear interface, and then complete the identification of modulation types, and most deep learning classification algorithms are is fragile and easy to be attacked, so when the signal classification model is a...

AI Technical Summary

Problems solved by technology

[0002] In daily life, the use of radio signals to transmit information is still the mainstream way of modern information transmission. After receiving radio signals, it is necessary to identify their specific modulation types, and then perform corresponding demodulation and extraction according to the modulation types. For the information carried by the signal, people usually need to spend a lot of time and energy to collect data samples of known categories, and use classification algorithms to train these data samples to form a clear interface, and then complete the identification of modulation types, and most deep learning classification algorithms are is fragile and easy to be attacked, so when the signal classification model is attacked by a specific attack algorithm, these radio signals will be subjected to some subtle perturbations so that the modulation type recognition model cannot correctly identify the modulation type of a given signal, so how to Selecting a suitable defense algorithm for defense has extremely important practical value

[0003] In the existing technology, there are many classification models, attack algorithms, defense algorithms, and detection algorithms in the field of radio signal modulation type classification. The current radio signal field does not correlate these information, let alone form a knowledge graph in this field. Radio In the attack and defense field of signal modulation type, only the classification accuracy after the attack is considered when selecting the attack and defense strategy, and other attributes are ignored, and the model and specific radio signal data are not analyzed. For many attack and defense detection algorithms, classification models and radio signal data, There is no unified representation method to associate them, and it is difficult to mine the implicit knowledge. For a classification model or radio signal, it is impossible to recommend a better attack and defense and detection strategy. Usually, only various algorithms can be tried.

Images