How the process of generating new candidate solutions in evolutionary search may evolve to become increasingly well-tuned to the target search space, particularly in the context of training of recurrent artificial neural networks to solve challenging reinforcement-learning-style problems.
|Research Interest Keywords|
|artificial intelligence, computational neuroscience, genetic algorithms, machine learning|
|The Robotics Institute is part of the School of Computer Science, Carnegie Mellon University.|
Contact Us | Update Instructions