Spring has arrived, and even AI has begun to lose weight...

Today's AI has acted on every aspect of life, but useful AI is also "chubby" AI most of the time. Relying on a large amount of training data and a large number of computing resources, it also requires the cooperation of many developers. This is an unbearable cost for industries with poor data samples, or for companies with insufficient funds. When technology is deployed, the process is also more cumbersome and slow, and it is difficult to meet the needs of agile innovation in enterprises. In order to enter more fields to shine, AI began the "slimming cause". Zero-sample learning, self-supervised learning, data augmentation algorithms... How will these technologies help AI slim down? How can ai that has slimmed down enter more fields?