What is a Teaching Machine?
- Understanding - A Teaching Machine should be able to interpret formulae, text, images, and diagrams the way humans do. They should be able to “read”, “see”, and “understand” a problem, decipher what is given, what is asked, and what background knowledge is assumed in the context of the problem.
- Thinking - A Teaching Machine should be able to consume, represent, and explore the “knowledge” it is given to discover one or more solutions to the problem. It should be able to differentiate a “good” solution from a “bad” solution and eventually become “smarter” at problem solving with more practice.
- Personalization - A Teaching Machine should be able to assess at a very fine grained level what a student knows, predict what s/he is ready to learn next, what is the best content to show to this student based on his/her learning style, and what sequence of problems the student should practice so s/he learns the concept most efficiently.
- Communication - A Teaching Machine should be able to have a conversation with the student beyond just question answers, it should also be able to recognize any form of nonverbal, implicit communication such as emotions and gestures to address the emotional, motivational, and intellectual aspects of the learning process.