Log in or register to comment What is a concept students (can / should / shouldn’t) learn from the section we read? At what age? In what context? (school, afterschool, course, in connection to __ science concept?) AI Book Club Discussions Comments I think students can and should learn the difference between symbolic and sub-symbolic AI, even if they are not interested in the detailed history of their development and periods of winter/spring. This could fit into a high school CS intro class. There are also plenty of elements of math in this chapter that could fit into the relevant high school math class. Finally, this chapter suggests (I hope a later chapter deals with) issues with bias in machine learning, which would be interesting to explore in a high school civics or psychology class. Learning the history of AI development can show students that there is often more than one promising way to reach a goal. Different teams took different approaches, and they both made different important contributions. I also agree that having discussion about the implications of new technology is really important. We don't know all of the ways that AI will impact our lives. And, students have a role to play in this. It is really important to understand how this works so that they can be informed. I think we'll get to Ethics in chapter 7. There are some really good student activities on the topic of Ethics and AI. In designing a GUTS AI curriculum, I'm thinking we want to put Ethics up front and center.... your thoughts? (I think the framing/ organization of the new curriculum is something we could discuss at length) --Irene In reply to sgibbs: I agree with you Sue! I found the explanation of symbolic and sub-symbolic approaches/methods fascinating, and think that it could prompt very interesting high school discussions regarding creativity in approaches to programming and programming architecture. I love that the "anarchy of methods" paves the way for "individuality" in use and approaches to computer language, as we draw parallels to human language. I agree with Susan that students can certainly get a good sense of expert systems Symbolic AI and sub symbolic AI . “A symbolic AI program’s knowledge consists of words or phrases (the “symbols”), typically understandable to a human, along with rules by which the program can combine and process these symbols in order to perform its assigned task.” Word story examples similar to the cannibals example provided allow students to see how to create a symbolic rule algorithm that is reusable and works. "Sub-symbolic AI is a stack of equations" ( I love this statement which to me emphasizes that it is absolutely complex) used to mimic the human brain giving weights to inputs , summing them to determine an output. The example provided by Irene at the AI conference with the image and weights applied to the image to determine what features were in the image is a good activity to explain how neural networks learn and form conclusions. A perfect example for students to decompose this abstract process into a way of building the knowledge of how non-symbolic AI works without having to know the stack of equations. In reply to Kelly Powers: We'll play that game next week (at least that is the plan). I'm hoping to use zoom chat where different people acting as nodes can pass messages to each other. --Irene Log in To Post New Content in the Forum.