GAN-Based Intrusion Detection System for IoT Services

With the great development of modern technology, the Internet of things are gradually changing the way we live. Many industry experts believe it will be the biggest technological shock since cloud computing because plenty of data can be collected, stored and analyzed than ever before. Thus, as an important technology, IoT services have been extensively used in not only our daily life but also Industry and agriculture. It is worth nothing that attackers are increasingly turning their attention to the public or personal IoT services. Hence, an effective detection scheme should be conducted as soon as possible. Generating Adversarial Network (GANs) can dispose the dataset of real life which is complex high-dimensional distribution, making it very suitable for the detection of abnormal state. However, at present, GANs commonly used for pattern recognition and speech recognition but rarely used in the field of information safety. In this paper, we used the an improved GANs model try to perform intrusion detection in IoT services. In the experiment, we improved the architecture of GANs to achieve the most advanced performance and shorten the training time when being compared with the present published other methods.