1 code implementation • 12 Aug 2021 • Sharan Vaswani, Olivier Bachem, Simone Totaro, Robert Mueller, Shivam Garg, Matthieu Geist, Marlos C. Machado, Pablo Samuel Castro, Nicolas Le Roux
Common policy gradient methods rely on the maximization of a sequence of surrogate functions.
no code implementations • 3 Jun 2021 • Lorenzo Steccanella, Simone Totaro, Anders Jonsson
In this paper we present a novel method for learning hierarchical representations of Markov decision processes.
no code implementations • 12 Nov 2020 • Lorenzo Steccanella, Simone Totaro, Damien Allonsius, Anders Jonsson
Sparse-reward domains are challenging for reinforcement learning algorithms since significant exploration is needed before encountering reward for the first time.
Efficient Exploration Hierarchical Reinforcement Learning +2
no code implementations • ICML Workshop LifelongML 2020 • Lorenzo Steccanella, Simone Totaro, Damien Allonsius, Anders Jonsson
Sparse-reward domains are challenging for reinforcement learning algorithms since significant exploration is needed before encountering reward for the first time.
Hierarchical Reinforcement Learning reinforcement-learning +1
1 code implementation • 16 May 2020 • Simone Totaro, Ioannis Boukas, Anders Jonsson, Bertrand Cornélusse
We propose a novel model based reinforcement learning algorithm that is able to address both types of changes.
Model-based Reinforcement Learning reinforcement-learning +1
no code implementations • 13 May 2020 • Dominik Thalmeier, Hilbert J. Kappen, Simone Totaro, Vicenç Gómez
We identify PICE as the infinite smoothing limit of such technique and show that the sample efficiency problems that PICE suffers disappear for finite levels of smoothing.
no code implementations • 11 Jul 2018 • Simone Scardapane, Steven Van Vaerenbergh, Danilo Comminiello, Simone Totaro, Aurelio Uncini
Gated recurrent neural networks have achieved remarkable results in the analysis of sequential data.
2 code implementations • 13 Jul 2017 • Simone Scardapane, Steven Van Vaerenbergh, Simone Totaro, Aurelio Uncini
Neural networks are generally built by interleaving (adaptable) linear layers with (fixed) nonlinear activation functions.