Epidemic prevention robot knowledge learning and transfer method and system

Abstract

The present invention provides a method and system for knowledge learning and transfer of epidemic prevention robots, including: step 1: according to the epidemic prevention scene applied by the robot, constructing a subtask offline knowledge base based on manual and prior knowledge; step 2: constructing according to the offline knowledge base Based on the multi-task sub-task network model, and perform task transfer learning training; Step 3: Build a knowledge reasoning engine based on the sub-task network model and task transfer learning training; Step 4: Build a knowledge map based on the knowledge reasoning engine; Step 5: Based on the knowledge The graph performs pattern training based on the task environment; Step 6: Carry out knowledge transfer based on the pattern training results based on the task environment. The invention allows robots to be used more widely in epidemic prevention and control, thereby liberating manpower and reducing the risk of staff being infected during epidemic prevention and control, and can play an important role in epidemic prevention and control.

Description

technical field [0001] The present invention relates to the technical field of epidemic prevention robots, in particular to a method and system for knowledge learning and transfer of epidemic prevention robots. Background technique [0002] Robots are often used to perform dangerous, repetitive tasks to free up humans. Just enough to meet the requirements for epidemic prevention. Therefore, more and more places are considering the use of epidemic prevention robots for epidemic prevention and control. For the anti-epidemic robot, its ability to learn and update continuously through autonomous learning is very necessary, but the traditional convolutional neural network has insufficient training samples, high training overhead, and few sample labels under the condition of small samples. It is difficult to apply in the epidemic prevention robot knowledge learning and transfer method proposed in this paper, which can just solve this problem. [0003] Patent document CN10974074...

AI Technical Summary

Problems solved by technology

For the anti-epidemic robot, its ability to learn and update continuously through autonomous learning is very necessary, but the traditional convolutional neural network has insufficient training samples, high training overhead, and few sample labels under the condition of small samples. It is difficult to apply in the epidemic prevention robot knowledge learning and transfer method proposed in this paper can just solve this problem

Images