Computer-implemented method for determining an adversarial patch for a machine learning system (60), wherein the machine learning system (60) is configured for image analysis and determines an output signal (y,yi) based on an input image (x, xi), wherein the output signal (y, yi) is determined based on an output of an attention layer (l) of the machine learning system (60), wherein the adversarial patch is determined by optimizing the adversarial patch with respect to a loss function, wherein the loss function comprises a term Lkq1lLkq2lLkq3l that characterizes a sum of attention weights of the attention layer (l) with respect to a position of the adversarial patch in the input image (x, xi) and the method comprises a step of maximizing the term.
DEVICE AND METHOD FOR DETERMINING ADVERSARIAL PATCHES FOR A MACHINE LEARNING SYSTEM
VORRICHTUNG UND VERFAHREN ZUR BESTIMMUNG VON KONTRADIKTORISCHEN PATCHES FÜR EIN MASCHINENLERNSYSTEM
DISPOSITIF ET PROCÉDÉ DE DÉTERMINATION DE PASTILLES CONTRADICTOIRES POUR UN SYSTÈME D'APPRENTISSAGE DE MACHINE
2023-08-16
Patent
Elektronische Ressource
Englisch
IPC: | G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / B25J Manipulatoren , MANIPULATORS / B60W CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION , Gemeinsame Steuerung oder Regelung von Fahrzeug-Unteraggregaten verschiedenen Typs oder verschiedener Funktion / G05B Steuer- oder Regelsysteme allgemein , CONTROL OR REGULATING SYSTEMS IN GENERAL / G06T Bilddatenverarbeitung oder Bilddatenerzeugung allgemein , IMAGE DATA PROCESSING OR GENERATION, IN GENERAL / G06V |
Determining strength of riveted patches
Engineering Index Backfile | 1937
|Online Contents | 2000
Online Contents | 2011