Learning Compositional Neural Programs with Recursive Tree Search and Planning

Thomas Pierrot
Thomas Pierrot
Guillaume Ligner
Guillaume Ligner
Alexandre Laterre
Alexandre Laterre
David Kas
David Kas
Karim Beguir
Karim Beguir

ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 32 (NIPS 2019), pp. 14646-14656, 2019.

Cited by: 7|Bibtex|Views128|Links
EI
Keywords:
reinforcement learningsample complexitytower of hanoi

Abstract:

We propose a novel reinforcement learning algorithm, AlphaNPI, that incorporates the strengths of Neural Programmer-Interpreters (NPI) and AlphaZero. NPI contributes structural biases in the form of modularity, hierarchy and recursion, which are helpful to reduce sample complexity, improve generalization and increase interpretability. Alp...More

Code:

Data:

Your rating :
0

 

Tags
Comments