Data augmentation is a creative process that involves manipulating data. It involves creating artificial data that is similar, but not identical, to the data you are working with. This allows the model to learn to recognize patterns in the augmented dataset instead of simply memorizing the original dataset. It was first introduced in 1974 by Frank Rosenblatt, a pioneer in AI, as a way to create artificial data to increase the size of a dataset in order to train a model. It can also be used to prevent overfitting, which can cause AI to lose its ability to generalize.
AI, blockchain, bioengineering
Create your free account to unlock your custom reading experience.
Continue reading: https://hackernoon.com/how-will-data-augmentation-affect-machine-learning?source=rss