Simbrain Workshop on July, 04. - 06.

In this 3-day course we will explore the basic principles of neural networks and their relationship to cognitive science and education. No prior experience with neural networks or mathematical training will be assumed. The workshop will be hands-on: students will build numerous neural network simulations during the course using a visually-oriented simulation package, Simbrain. In addition to learning about neural networks, at a meta-level students will learn about how to use Simbrain to  teach neural network concepts to a broader range of students than has traditionally been possible (students as young as 10-12, or even younger, should be able to learn the basic concepts using this software).

Schedule

• 04.07. (WED)  09:00 - 15:00 (in 3220)
• 05.07. (THU)  09:00 - 15:00 (in 3220)
• 06.07. (FRI)  09:00 - 15:00 (in RL13a-CIP Personalmensa)

Topics

Topics to be covered: bui lding neural networks, plot ting network activity, studying network dynamics, associative memory, recurrent networks, supervised  learning and back-propagation, unsupervised learning, memories as attractors of a dynamical system, using a network to control an agent in an environment using a neural network.  Advanced topics to be covered, time permitting: spiking networks, genetic algorithms, reinforcement learning, echo state networks.

Bio

Jeff Yoshimi is an Associate Professor of phi losophy and cognitive science at the University of California,  Merced, a new university which opened in 2005, and the f irst  new campus of the University of California System bui lt since the 1960's. His research is interdiscipl inary, integrating methods from philosophy, cognitive science, and computational modeling. He received his BA in philosophy from UC Berkeley, his Ph.D. in philosophy from UC Irvine, and was a Postdoctoral Fellow at UC San Diego, before he joined and helped to found UC Merced in 2004.