I will start teaching a class in artificial intelligence starting next week. While preparing this class, I found out how my thoughts differ from what is written in most of the book. Therefore the concept of the class will be different. It will be a mix between a technical class and a more philosophical way about the way I define some of the terms for myself and the consequences.
The first and foremost question is: What is intelligence. And many people see intelligence related to humans. As a consequence we will have to look at humans with all the things that are working excellent in humans and why we deceive ourselfs and others. We have to cover how we view the world, that humans only see what they „want to see“.
The class is difficult, because the terms are used quite different for humans and computers. When technicians talk about neural networks, they usually have a stack of nonlinear elements that is trained by backpropagation. In contrast to this very simplified, the human has a different structure. We have highly cross-linked neurons, that have a temporal dependent pattern in which they fire. These neurons can be organized in columns. Chemicals influence the way they fire. After firing they need a certain time to recuperate and to build up new potential to fire. The closest similarity between neural networks in computer science and humans are the name.
I am really looking forward to this class and hope that the students will like to go on this journey with me.
0 Comments