Google Southwest Regional Alliance Teacher training

Day 1

On May 11

I was fortunate to attend a training session held by Google in Chongqing Institute of Science and Technology. I feel very happy, because for us this kind of opportunity is very rare. Therefore, I got what I needed when I arrived on the afternoon of 11th, visited the hotel and the school, and carefully read the learning tasks for these two days.

The second day

On May 12

Group photo

Most of them are students from famous universities in southwest China, all of them come to learn technology. I feel very happy to learn with them, communicate with them and learn their learning experience.

The opening lecture

It mainly explains the basic process of these two days and the knowledge that we need to learn, as well as the tasks that teachers need to learn.

Tensorflow learning

This is the first part of the training, which impressed me deeply. At the beginning, we started with simple machine learning and neural network introduction and the use of Tensorflow. After the teacher’s explanation, we digested the class contents explained by him. After a brief digestion, we carried out our own operation practice.

Android+ Tensorflow+OpenCV

“Android+ Tensorflow+OpenCV” this is a transfer learning based on Tensorflow. When these three applications are combined sensory neural network based implementation is very difficult,

Especially in the training of neural network neurons need to use the mathematical knowledge of normal distribution. So from here again to understand that to do an excellent programmer must have a good mathematical foundation. When you train neurons, it’s getting better, but it’s still not 100%, so no machine can be completely error-free.

The third day

On May 13

The car assembly

One of the most interesting is the assembly of the car. The team worked together to assemble a car using the parts used to assemble two cars, because one of the car chips burned out.

Tensorflow, OpenCV, Android Studio

“Development of identification apps such as Tensorflow, OpenCV and Android Studio for traffic signs”, in which we mainly realize the automatic driving of cars. It is based on Google’s Tensorflow Object Detection API implementation.