That the generator has learned how to

RABIULISLAMSP88

simulate the distribution of real data and the discriminator can no longer provide useful feedback to guide the training of the generator. As an example, suppose there is a counterfeiter generator and a police discriminator. They are engaging in a "cat and mouse" game. The counterfeiter's goal is to create counterfeit bills that are as authentic as possible to deceive the police. He may only be able to produce crude counterfeit bills at first but as time goes by his skills gradually improve. Improvement can produce more and more realistic counterfeit banknotes. This is like the generator can only generate data that is


quite different from the real data at the beginning, but as the training progresses, the generator's generation ability gradually improves and can generate data that is closer and closer to the real data. The police's goal is to distinguish real banknotes from counterfeit banknotes as accurately as possible. He may have a weak ability to identify counterfeit banknotes at the beginning, but as he studies counterfeit banknotes, his recognition ability gradually improves and he Rich People Phone Number List can identify counterfeit banknotes more accurately. . This is like the discriminator can only roughly distinguish between real data and generated data at the





beginning, but as the training progresses, the discriminant ability of the discriminator gradually improves and can more accurately distinguish between real data and generated data. In this process, the counterfeit banknote maker and The police are constantly improving their skills and eventually reaching a dynamic balance. It's like the generator and the discriminator are constantly improving their capabilities during the training process. In the end, we can harvest very realistic counterfeit banknotes "generators and very capable police officers at the same time." "Discriminator. 2. Application scenarios N is widely used in many fields. The following are some specific examples. Image generation N can generate high-quality images such as eer, eere, etc. For example, training an N that learns a