Generative adversarial sample processing system of natural language model

The invention provides a generative adversarial sample processing system of a natural language model, which relates to the field of electrical digital data processing and comprises a model docking module, an adversarial strategy processing module, an adversarial sample generation module, a sample evaluation module and an iterative optimization module. The model docking module is used for docking with a natural language model and receiving a training sample of the model, the adversarial strategy processing module sets adversarial strategy information, and the adversarial sample generation module processes the training sample according to an adversarial strategy to generate an adversarial sample; the sample evaluation module is used for operating adversarial samples in the model and screening out adversarial samples with insufficient adversarial resistance, and the iterative optimization module is used for performing iterative processing on the screened adversarial samples to improve adversarial resistance; acc