KLay: Accelerating Arithmetic Circuits for Neurosymbolic AI

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Maene, Jaron, Derkinderen, Vincent, Martires, Pedro Zuidberg Dos
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866929723325546496
author Maene, Jaron
Derkinderen, Vincent
Martires, Pedro Zuidberg Dos
author_facet Maene, Jaron
Derkinderen, Vincent
Martires, Pedro Zuidberg Dos
contents A popular approach to neurosymbolic AI involves mapping logic formulas to arithmetic circuits (computation graphs consisting of sums and products) and passing the outputs of a neural network through these circuits. This approach enforces symbolic constraints onto a neural network in a principled and end-to-end differentiable way. Unfortunately, arithmetic circuits are challenging to run on modern AI accelerators as they exhibit a high degree of irregular sparsity. To address this limitation, we introduce knowledge layers (KLay), a new data structure to represent arithmetic circuits that can be efficiently parallelized on GPUs. Moreover, we contribute two algorithms used in the translation of traditional circuit representations to KLay and a further algorithm that exploits parallelization opportunities during circuit evaluations. We empirically show that KLay achieves speedups of multiple orders of magnitude over the state of the art, thereby paving the way towards scaling neurosymbolic AI to larger real-world applications.
format Preprint
id arxiv_https___arxiv_org_abs_2410_11415
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle KLay: Accelerating Arithmetic Circuits for Neurosymbolic AI
Maene, Jaron
Derkinderen, Vincent
Martires, Pedro Zuidberg Dos
Machine Learning
A popular approach to neurosymbolic AI involves mapping logic formulas to arithmetic circuits (computation graphs consisting of sums and products) and passing the outputs of a neural network through these circuits. This approach enforces symbolic constraints onto a neural network in a principled and end-to-end differentiable way. Unfortunately, arithmetic circuits are challenging to run on modern AI accelerators as they exhibit a high degree of irregular sparsity. To address this limitation, we introduce knowledge layers (KLay), a new data structure to represent arithmetic circuits that can be efficiently parallelized on GPUs. Moreover, we contribute two algorithms used in the translation of traditional circuit representations to KLay and a further algorithm that exploits parallelization opportunities during circuit evaluations. We empirically show that KLay achieves speedups of multiple orders of magnitude over the state of the art, thereby paving the way towards scaling neurosymbolic AI to larger real-world applications.
title KLay: Accelerating Arithmetic Circuits for Neurosymbolic AI
topic Machine Learning
url https://arxiv.org/abs/2410.11415