While performing experiments in lab you refer lab manual or teacher instructs you. After what is to be done and how you perform your experiment on your own. This is known as learning by advice.

Learning by advice requires more inference then rote learning. There are two main approaches for advice taking:

1. Automating all aspects of advice taking: In this method agent takes abstract high level advice from the teacher and then converts this advice into protocols which then help in guiding performance elements of the agent.

2. Using tools such as knowledge base editor and debuggers of advice taking: Here, an expert is a fundamental part of the learning agent. This type of advice taking is used to assist an expert to explain his expertise into detailed protocols. While storing protocols in the knowledge base, knowledge is transformed into an operational form.