Machine learning

Defination: the study of computer systems that improve their performance through experience.

• learn existing and known structures and rules
• discover new findings and structures

Supervised Learning:

• Goal: learn a mapping from inputs x to outputs y
• Training data: a labeled set of input-output pairs

人脸识别应用的实际逻辑：输入两张人脸，输出0 / 1，0代表两张人脸不一致，1代表两张人脸一致

自动驾驶：训练样本均为晴天拍摄，而实际开出去后为夜晚/恶劣天气时效果较差；永远会碰到可能数据中没见过的情况。解决方式：发现问题后补数据，进行增量训练

• 当y离散时——Classification
• 当y是实际数字时——Regression

应用：object detection / semantic segmentation

step1 is much more important

Unsupervised Learning:

• Goal: finding “interesting patterns”
• only given inputs