Cara Reedy, RPI Graduate Student

 

Cara Reedy, RPI Graduate Student

Sage 4101

February 22, 2017 12:00 PM - 1:00 PM

Artificial life is a field that simulates life or life processes with the hopes of either gaining insight into real-life biology, or to explore forms of life that are not or could not be possible on Earth. It often uses artificial evolution and/or learning methods like neural networks. This talk will present an artificial life simulation with a simple environment and agents that use neural networks to make decisions and interact with other objects. The agents can learn over time through the outcomes of their actions, but the structure and the starting weights of their neural nets are controlled by a diploid genetics system. Because of this, both the weights and the structure of the networks that the agents use can evolve over time and generations. This simulation is also intended to eventually be deployed on neuromorphic computing hardware, which could both allow for a large number of agents in the simulation, and test the abilities and limitations of the hardware.

 

Add to calendar
Share|