Paul Schydlo

I'm fascinated by how the mind takes in information from our senses, specifically 2D retinal images, and extracts only what it needs to solve the task at hand. How it does so much with so little energy.

At the moment, I study how we represent this information and how we can apply some of the same principles in our perception algorithms.

Application scenarios cover the fields of human action prediction, intuitive physics, planning and other related prediction tasks.

Currently, I'm pursuing my Ph.D in joint supervision by prof. Katerina Fragkiadaki at the Machine Learning Department at CMU in Pittsburgh and prof. José Santos-Victor at the Institute for Systems and Robotics in Lisbon.

"What I cannot create, I do not understand." - R. Feynman

or at:

Research Interests


Anticipation in Human-Robot Cooperation: A Recurrent Neural Network Approach for Multiple Action Sequences Prediction
Paul Schydlo, Mirko Rakovic, Lorenzo Jamone, José Santos-Victor
International Conference on Robotics and Automation (ICRA) 2018

Personal Projects

Teaching Experience


Research Projects



2019 - Present Carnegie Mellon University, Pittsburgh
Machine Learning Department, Ph.D Co-Supervision by prof. Katerina Fragkiadaki.
2019 - Present Instituto Superior Técnico, Lisbon, Portugal
Ph.D. Electrical and Computer Engineering, Supervised by prof. José Santos-Victor and prof. Katerina Fragkiadaki from CMU
2015 Technical University Delft, Netherlands
M.Sc. Systems and Control, Erasmus exchange semester
2012-2017 Instituto Superior Técnico, Lisbon, Portugal
Integrated M.Sc. Electrical and Computer Engineering, major in Systems, Decision and Control