Neural networks are not a new method, the first artificial neural network was devised in 1943, but advances in computational power and speed have made them a much more viable strategy for solving complex problems over the last 5-10 years. Originally devised by mathmaticians and neuroscientists to illustrate the fundamental principles of how brains might work they lost favor in the second half of the 20th century only to surge in popularity in the 20-teens as software engineers used them to resolve mathmatically intractable problems. The application of neural networks to learning problems has been ongoing for 20 years, often to predict student behvior or to parse unstructured data such as student writing samples and provide natural sounding feedback through AI avatars.
- Be able to explain the utility of artificial neural networks
- Be able to explain the backpropagation algorithm
- Be able to build a basic neural network to solve a prediction problem
In this unit you will be building a neural network to predict student attentional state from webcam imags. As background to this task please read over the follwing materials and watch the methodological videos. If you find any other useful materials please add them under Additional Materials at the end of the this page and pull request the change back to this repo.
Sanderson, G. (2017). But what is a Neural Network? 3Blue1Brown.
Bling, S. (2017). MariFlow - Self-Driving Mario Kart with Recurrent Neural Network
Nielsen, M. (2015). Neural Networks & Deep Learning. Determination Press:San Francisco, CA
Chapter 1
Charpter 2
For more detail:
Stergiou, C. & Siganos, D. (2000). Neural Networks.
Hartnett, K. (2019). Foundations Built for a General Theory of Neural Networks
Once you have completed all tasks in the unit, please complete the knowledge check.
Lewis-Kraus, G. (2016). The Great AI Awakening. The New York Times: New York, NY
Roberts, E. (2000). History in Neural Networks. Stanford University: Palo Alto, CA