Differentiable Inductive Logic Programming in High-Dimensional Space [article]

Stanisław J. Purgał, David M. Cerna, Cezary Kaliszyk
2022 arXiv   pre-print
Synthesizing large logic programs through Inductive Logic Programming (ILP) typically requires intermediate definitions. However, cluttering the hypothesis space with intensional predicates often degrades performance. In contrast, gradient descent provides an efficient way to find solutions within such high-dimensional spaces. Neuro-symbolic ILP approaches have not fully exploited this so far. We propose an approach to ILP-based synthesis benefiting from large-scale predicate invention
more » ... g the efficacy of high-dimensional gradient descent. We find symbolic solutions containing upwards of ten auxiliary definitions. This is beyond the achievements of existing neuro-symbolic ILP systems, thus constituting a milestone in the field.
arXiv:2208.06652v1 fatcat:llteibxvrjeolpn6zqrruu3yym