KINS: Knowledge Injection via Network Structuring


Matteo Magnini, Giovanni Ciatto, Andrea Omicini

Proceedings of the  37th Italian Conference on Computational Logic
 2022

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

(keywords) explainable AI; symbolic knowledge injection; KINS; PSyKI
 @inproceedings{kins-cilc2022,
keywords = {explainable AI; symbolic knowledge injection; KINS; PSyKI},
year = 2022,
author = {Magnini, Matteo and Ciatto, Giovanni  and Omicini, Andrea},
title = {KINS: Knowledge Injection via Network Structuring},
booktitle = {Proceedings of the  37th Italian Conference on Computational Logic}

Events

  • 37th Italian Conference on Computational Logic (CILC 2022) — 29/06/2022–01/07/2022

Publication

— authors

Matteo Magnini, Giovanni Ciatto, Andrea Omicini

— status

camera-ready sent

— sort

paper in proceedings

Venue

— volume

Proceedings of the  37th Italian Conference on Computational Logic

— publication date

2022

BibTeX

— BibTeX ID
kins-cilc2022
— BibTeX category
inproceedings

View this PDF full screen

You do not have the plugin required to display this PDF file. You can still download it: extraamas-2021-iter.pdf

Partita IVA: 01131710376 - Copyright © 2008-2022 APICe@DISI Research Group - PRIVACY