A view to a KILL: Knowledge Injection via Lambda Layer

   Info_icon.svg   dati    Eye_icon.svg   allegati    BibTeX_logo.svg   bibtex
Angelo Ferrando, Viviana Mascardi (a cura di)
WOA 2022 – 23rd Workshop “From Objects to Agents”, pp. 61–76
CEUR Workshop Proceedings (AIxIA Series) 3261
Sun SITE Central Europe, RWTH Aachen University
novembre 2022

We propose KILL (Knowledge Injection via Lambda Layer) as a novel method for the injection of symbolic knowledge into neural networks (NN) allowing data scientists to control what the network should (not) learn. Unlike other similar approaches, our method does not (i) require ground input formulae, (ii) impose any constraint on the NN undergoing injection, (iii) affect the loss function of the NN. Instead, it acts directly at the backpropagation level, by increasing penalty whenever the NN output is violating the injected knowledge. Experiments are reported to demonstrate the potential (and limits) of our approach.

parole chiave   symbolic knowledge injection, AI, ML, neural networks, KILL, PSyKI
presentazione di riferimento
evento origine
rivista o collana
Journal_icon.svg CEUR Workshop Proceedings (CEUR-WS.org)
progetto finanziatore
Project_icon.svg EXPECTATION — Personalized Explainable Artificial Intelligence for decentralized agents with heterogeneous knowledge (01/04/2021–31/03/2024)
funge da
pubblicazione di riferimento per presentazione