Welcome to my personal GitHub page!

I am a researcher in Reinforcement Learning and Control methodology. I have a Ph.D. in Electrical Engineering from KTH Royal Insitute of Technology (Stockholm), where I was supervised by Prof. Proutiere Alexandre at the Division of Decision and Control Systems.

My research involves the study of Reinforcement Learning and data-driven control, including

• Exploration in Reinforcement Learning
• Reinforcement learning methodology
• Conformal prediction
• Learning in control
• Poisoning attacks on Sequential Decision Makers (Reinforcement Learning, Adaptive Control, Data-Driven Control)
• Attack detectability

Check out also my website

• Reading list: here I keep track of interesting papers

• Research related

  • Model Free Active Exploration in Reinforcement learning
  • Conformal Prediction for Off-Policy Evaluation in MDPs
  • Data poisoning of linear systems
  • OSRL-SC, Optimal Representation Learning in Multi-Task Bandits
  • TZDDPC, Tube-based Zonotopic Data-Driven Predictive COntrol
  • STTMPC, Self-Tuning Tube-based MPC
  • Privacy Stochastic Systems
  • DPE, Optimal algorithms for multi-agent bandit models
  • Poisoning of Data Driven Controllers (TBA)
  • Optimal Attacks on RL Policies (TBA)

• Libraries

  • PyZPC: an implementation of ZPC in Python (Zonotopic Data Driven Predictive Control)
  • PyZonotope: a library that enables the usage of Zonotopes in CVXPY
  • PythonVRFT: a python implementation of the VRFT algorithm
  • PyDataDrivenReachabilityAnalysis: a python implementation of data-driven rechability analysis based on zonotopes
  • PyQuadraticFormNormal: Python library that implements the computation of the distribution of a linear combination of non-centered χ2 random variables
  • PyDeePC: A python implementation of the data-driven predictive control algorithm DeePC
  • py-lower-bound-bai: A python library that solves the best-arm identification problem in classical multi-armed bandit models.
  • nnGA: Neural Network Genetic Algorithm library used for deep learning problems
  • PytorchRBFLayer: Pytorch RBF Layer implements a radial basis function layer in Pytorch. Radial Basis networks can be used to approximate functions.
  • AdaptiveControlLibrary: Adaptive Control Library in Matlab
  • TBC

• Other repos

  • Algorand Royalty Fees: a tutorial that shows how to implement royalty fees on the Algorand blockchain
  • TBC