Skip to yearly menu bar Skip to main content


Spotlight Poster

Compositional Generalization Across Distributional Shifts with Sparse Tree Operations

Paul Soulos · Henry Conklin · Mattia Opper · Paul Smolensky · Jianfeng Gao · Roland Fernandez

[ ] [ Project Page ]
Wed 11 Dec 11 a.m. PST — 2 p.m. PST

Abstract: Neural networks continue to struggle with compositional generalization, and this issue is exacerbated by a lack of massive pre-training. One successful approach for developing neural systems which exhibit human-like compositional generalization is $\textit{hybrid}$ neurosymbolic techniques. However, these techniques run into the core issues that plague symbolic approaches to AI: scalability and flexibility. The reason for this failure is that at their core, hybrid neurosymbolic models perform symbolic computation and relegate the scalable and flexible neural computation to parameterizing a symbolic system. We investigate a $\textit{unified}$ neurosymbolic system where transformations in the network can be interpreted simultaneously as both symbolic and neural computation. We extend a unified neurosymbolic architecture called the Differentiable Tree Machine in two central ways. First, we significantly increase the model’s efficiency through the use of sparse vector representations of symbolic structures. Second, we enable its application beyond the restricted set of tree2tree problems to the more general class of seq2seq problems. The improved model retains its prior generalization capabilities and, since there is a fully neural path through the network, avoids the pitfalls of other neurosymbolic techniques that elevate symbolic computation over neural computation.

Live content is unavailable. Log in and register to view live content