Fernando Gama

Postdoctoral Research Associate

Rice University•  August 2021 - Present

Research on the theory of graph neural networks, as it pertains to guarantees on transferability. Applications to telecommunications. Research on particle filtering and signal processing.

Postdoctoral Scholar

University of California, Berkeley•  September 2020 - August 2021

Research on graph neural networks for distributed control, including guarantees on stability and robustness. Applications to power grids. Research on the theory of graph neural networks via graph filtering and graph frequency response.

Research Fellow

University of Pennsylvania•  August 2014 - August 2020

Research on graph neural networks leveraging graph signal processing. Development of novel architectures and establishment of theoretical guarantees on stability and robustness. Pioneering applications to robotics, sensor networks and power grids. Additional research on ergodicity of random stationary graph processes, sampling and sketching of random graph signals, and overlapping hierarchical clustering of network data.

Research Intern

Facebook•  August 2018 - December 2018

Research on generative models of graph data using generative adversarial networks (GANs) and recurrent neural networks (in particular, LSTMs).

Visiting Researcher

TU Delft•  July 2017 - September 2017

Research on graph neural network architectures. Additional research on controllability of bandlimited graph signals.