A view to a KILL: Knowledge Injection via Lambda Layer


Matteo Magnini  /  Matteo Magnini, Giovanni Ciatto, Andrea Omicini

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.

Events

  • 23rd Workshop “From Objects to Agents” (WOA 2022) — 01/09/2022–03/09/2022

Publications

Talk

— speakers

— authors

— sort

talk

— language

wgb.gif

Context

23rd Workshop “From Objects to Agents” (WOA22)

— where

Genova, Italy

— when

01/09/2022

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