Password security intelligent detection method based on de-noising diffusion probability model

The invention provides a password security intelligent detection method based on a denoising diffusion probability model. The method comprises the following steps: step 1, randomly dividing a target password set into a password training set and a password test set in proportion; 2, constructing a de-noising diffusion probability model, and training the de-noising diffusion probability model by adopting a password training set; 3, generating a password dictionary by adopting the trained de-noising diffusion probability model according to user requirements; step 4, de-duplicating the generated password dictionary; step 5, calculating a collision rate of the password dictionary subjected to duplicate removal and the password test set; 6, if the collision rate is larger than a set threshold value, it is considered that potential safety hazards exist in the target password set, and otherwise, it is considered that the target password set meets the safety requirement. 本发明提供一种基于去噪扩散概率模型的口令安全智能检测方法。该方法包括：步骤1：将目标口令集按比