Strategies for Safety Goal Decomposition for Neural Networks

Faculty/Professorship: Cognitive Systems  
Author(s): Schwalbe, Gesina ; Schels, Martin
Corporate Body: ACM Computer Science in Cars Symposium (CSCS), 3, 2019, Kaiserslautern
Publisher Information: Bamberg : Otto-Friedrich-Universität
Year of publication: 2020
Pages: 3
Source/Other editions: 3rd ACM Computer Science in Cars Symposium – Future Challenges in Artificial Intelligence & Security for Autonomous Vehicles / German Chapter of the ACM e.V. : München, S. 3
Language(s): English
DOI: 10.20378/irb-47274
Licence: Creative Commons - CC BY-NC-SA - Attribution - NonCommercial - ShareAlike 4.0 International 
URN: urn:nbn:de:bvb:473-irb-472744
Neural networks (NNs) have become a key technology for solving highly complex tasks, and require integration into future safety argumentations. New safety relevant aspects introduced by NN based algorithms are: representativity of test cases, robustness, inner representation and logic, and failure detection for NNs. In this paper, a general argumentation structure for safety cases respecting these four aspects is proposed together with possible sources of evidence.
GND Keywords: Neuronales Netz; Beweisführung; Dekomposition; Sicherheit
Keywords: neural network, safety argumentation, safety goal decomposition
DDC Classification: 004 Computer science  
RVK Classification: ST 301   
Type: Conferenceobject
Release Date: 3. July 2020

File Description SizeFormat  
fisba47274.pdf124.89 kBPDFView/Open