Back to projects

Compositional Program Induction

Tackling discrete compositional program induction with latent structures

Program InductionLatticeCompositional Reasoning

This project explores how geometric deep learning principles can be applied to program induction tasks. The key insight is that programs have inherent symmetries that neural networks can exploit.

visit the project website