Practical Research on Primary Mathematics Teaching Based on Deep Learning

Table 1

The correspondence of deep learning in the two domains.

	Deep learning in computers	Deep learning in teaching science
Model	Input layer (initial data)	Challenging learning tasks
	Hidden layer (transmitting between neurons in the brain, processing information, contains multiple layers of neurons)	The existing knowledge in the cognitive structure
	Output layer (output-processed information)	The essence of mathematics
	Inspection model	Persistent evaluation
Goal	Get the machine to be able to learn	Promote students to have the ability of deep learning, independently explore the connection between knowledge, from “learn” to “learn” can “learn”
Method	Use a large amount of training data to determine the model	Under the guidance of teachers, students complete tasks, activate more knowledge elements, and independently explore the relationship between knowledge elements
Keynote	Training process, neuronal activation, and determination of the weights	Students’ thinking process, the activation of knowledge element and the relationship between knowledge element