Algorithm 2 Adversarial training.
Input:
ori_IDS: The pretrained IDS,  OTD: The original training datasets, num_adv_samples,  adv_samples
Output:
opt_IDS: The optimization of IDS;
size = length(adversarial_samples); if size > num_adv_samples:  part_samples = OTD.take(2*num_adv_samples);  balance_samples = part_samples.contact(adv_samples);  opt_IDS = ori_IDS.train(balance_samples); return opt_IDS; end if