Learn about the lecture
Fast Weights
Unlike the real brain, nearly all artificial neural networks have only two time scales: Neural activities change rapidly and the weights of connections between neurons change slowly. Many interesting computational properties can be implemented by adding an overlay of weights that adapt rapidly and decay rapidly. Dr. Hinton will discuss some of these properties, including accessing old memories, implementing true recursion, and the Poincaré effect.
Register to attend
To join us, register by Tuesday, March 1. Prior to the event, you will receive a Zoom login link in a confirmation email.
You will have an opportunity to indicate any accessibility requirements as you register. If you have any questions, contact
ULS@andrew.cmu.edu.
Carnegie Mellon University programs and events are open to all alumni, regardless of race, color, national origin, sex, handicap or disability, age, sexual orientation, gender identity, religion, creed, ancestry, belief, veteran status or genetic information.