Muhammed Fadera is a PhD student at the Department of Mathematics and Statistics. His interest lies at the interest intersection of Machine learning, nonlinear dynamics and nonautonomous dynamical systems. His research has used tools from these disciplines to understand and interpret how RNNs make decisions. From dynamical system perspective, this involve understanding the nature of invariant objects like equillibria and periodic orbits that emerged during training and their subsequent interaction once input is added. Muhammed has also looked at ways to design machine learning models to solve finite state computation problems with minimal number of neurons. He is currently working on the fractal dimension of the subset of state space explored by RNNs once they have been successfully trained to perform a given task.