We propose a novel method to inject symbolic knowledge in form of Datalog formulæ into neural networks (NN), called Knowledge Injection via Network Structuring (KINS). The idea behind our method is to extend NN internal structure with ad-hoc layers built out of the injected symbolic knowledge. KINS does not constrain NN to any specific architecture, neither requires logic formulæ to be ground. Moreover, it is robust w.r.t. both lack of data and imperfect/incomplete knowledge. Experiments are reported, involving multiple datasets and predictor types, to demonstrate how KINS can significantly improve the predictive performance of the neural networks it is applied to.

Matteo Magnini, G.C. (2023). Knowledge injection of Datalog rules via Neural Network Structuring with KINS. JOURNAL OF LOGIC AND COMPUTATION, 33(8), 1832-1850 [10.1093/logcom/exad037].

Knowledge injection of Datalog rules via Neural Network Structuring with KINS

Matteo Magnini;Giovanni Ciatto;Andrea Omicini
2023

Abstract

We propose a novel method to inject symbolic knowledge in form of Datalog formulæ into neural networks (NN), called Knowledge Injection via Network Structuring (KINS). The idea behind our method is to extend NN internal structure with ad-hoc layers built out of the injected symbolic knowledge. KINS does not constrain NN to any specific architecture, neither requires logic formulæ to be ground. Moreover, it is robust w.r.t. both lack of data and imperfect/incomplete knowledge. Experiments are reported, involving multiple datasets and predictor types, to demonstrate how KINS can significantly improve the predictive performance of the neural networks it is applied to.
2023
Matteo Magnini, G.C. (2023). Knowledge injection of Datalog rules via Neural Network Structuring with KINS. JOURNAL OF LOGIC AND COMPUTATION, 33(8), 1832-1850 [10.1093/logcom/exad037].
Matteo Magnini, Giovanni Ciatto, Andrea Omicini
File in questo prodotto:
Eventuali allegati, non sono esposti

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11585/950567
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? 0
social impact