Feryal Behbahani
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Feryal Behbahani has a total of 6 co-authors. Here are the top co-authors by number of shared publications:
- Alexandre Laterre
Sorbonne Universite Sorbonne Universite
1 shared publication - Karim Beguir
InstaDeep
1 shared publication - Nando de Freitas
University of Washington
1 shared publication - Nicolas Perrin
CNRS Sorbonne Université Sorbonne Université
1 shared publication - Olivier Sigaud
Sorbonne Université, CNRS, Institut des Systèmes Intelligents et de Robotique (ISIR), Paris, France
1 shared publication - Thomas Pierrot
InstaDeep
1 shared publication
Latest Publications
- Open Access
Learning Compositional Neural Programs for Continuous Control
We propose a novel solution to challenging sparse-reward, continuous control problems that require hierarchical planning at multiple levels of abstraction. Our solution, dubbed AlphaNPI-X, involves three separate stages of learning. First, we use off-policy reinforcement learning algorithms with experience replay to learn a set of atomic goal-conditioned policies, which can be easily repurposed for many tasks. Second, we learn self-models describing the effect of the atomic policies on the environment. Third, the self-models are harnessed to learn recursive compositional programs with multiple levels of abstraction. The key insight is that the self-models enable planning by imagination, obviating the need for interaction with the world when learning higher-level compositional programs. To accomplish the third stage of learning, we extend the AlphaNPI algorithm, which applies AlphaZero to learn recursive neural programmer-interpreters. We empirically show that AlphaNPI-X can effectively learn to tackle challenging sparse manipulation tasks, such as stacking multiple blocks, where powerful model-free baselines fail.
Open AccessLearning Compositional Neural Programs for Continuous Control
We propose a novel solution to challenging sparse-reward, continuous control problems that require hierarchical planning at multiple levels of abstraction. Our solution, dubbed AlphaNPI-X, involves three separate stages of learning. First, we use off-policy reinforcement learning algorithms with experience replay to learn a set of atomic goal-conditioned policies, which can be easily repurposed for many tasks. Second, we learn self-models describing the effect of the atomic policies on the environment. Third, the self-models are harnessed to learn recursive compositional programs with multiple levels of abstraction. The key insight is that the self-models enable planning by imagination, obviating the need for interaction with the world when learning higher-level compositional programs. To accomplish the third stage of learning, we extend the AlphaNPI algorithm, which applies AlphaZero to learn recursive neural programmer-interpreters. We empirically show that AlphaNPI-X can effectively learn to tackle challenging sparse manipulation tasks, such as stacking multiple blocks, where powerful model-free baselines fail.