1. Study notes
Video learning content this week: https://www.bilibili.com/video/BV1Tb411H7uC?p=2
1) P2 probability theory and Bayesian prior
2) P3 matrix and linear algebra
2) "Gradient", "Gradient Descent" and "Bayes' Theorem"
Learning reference materials: https://blog.csdn.net/qq_20412595/article/details/81409744
gradient:
Baidu said that the gradient is a vector, which means that the directional derivative of a function at that point gets the maximum value along the direction, that is, the function changes the fastest along the direction (the direction of the gradient) at the point, the rate of change Maximum (the modulus of this gradient). In my understanding, for example, to climb the stairs to the second floor, no matter how the stairs are bent, the direction of the straight line between the two points is the fastest. This is the gradient, which is a vector, and every step of the stairs, each step rises. The fastest direction is the gradient that step is that point.
The partial derivatives of the variables of the multivariate function are written in the form of vectors, which is the gradient. Such as functions , its gradient
or
is
Gradient descent:
As the gradient mentioned above, the direction with the fastest and largest change is the gradient. Then, the gradient decreases according to a certain step in the opposite direction of the gradient, which is the gradient.
Bayes' theorem
The Bayesian formula is as follows:
Under the premise that condition B occurs, the probability of time A occurring is:
Bayes' theorem is about the probability that B occurs when the two events intersect, with the occurrence of A as the premise. The probability of encountering a crossroad on the road is 30%, and the probability of a right turn is 10%. If you want to turn right, the probability of turning right becomes 25% when you encounter a crossroad.