87
87c86153... 1y ago

Training neural network is using compute to find common patterns in input data. Its process of compressing what is common between each provided input segment into a generalized form. This generalized form can be then used to generate new data.

As opposed to popular view, where training AI on AI generated data is seen as providing worse output, the reality is opposite:

When we create generalization of some dataset, we reduce noise and extract what is important. When we train new AI on AI generated data, we generalize generalization, which results into a better generalization of original data.

Reply to this note

Please Login to reply.

Discussion

af
Condor 1y ago

Over generalization in psychology brings to abnormal unwanted behaviour. The line between healthy and sick generalization is very thin. The result anyway is not a better result but an useful generalized understanding of the external world that covers holes on detailed knowledge. Pretty good to predict big numbers of items behaviour,.very poor's predicting single items. We will see but remember that high IQ individuals that easily understand patterns are usually very weird human beings. Playing with fire...