Proceedings of the 1st International Conference on AI for People: Towards Sustainable AI, CAIP 2021, 20-24 November 2021, Bologna, Italy

Research Article

Towards Functional Safety Compliance of Recurrent Neural Networks

Download72 downloads
  • @INPROCEEDINGS{10.4108/eai.20-11-2021.2314139,
        author={Davide  Bacciu and Antonio  Carta and Daniele  Di Sarli and Claudio  Gallicchio and Vincenzo  Lomonaco and Salvatore  Petroni},
        title={Towards Functional Safety Compliance of Recurrent Neural Networks},
        proceedings={Proceedings of the 1st International Conference on AI for People: Towards Sustainable AI, CAIP 2021, 20-24 November 2021, Bologna, Italy},
        publisher={EAI},
        proceedings_a={CAIP},
        year={2021},
        month={12},
        keywords={functional safety dependability recurrent neural networks autonomous driving safety performance indicators},
        doi={10.4108/eai.20-11-2021.2314139}
    }
    
  • Davide Bacciu
    Antonio Carta
    Daniele Di Sarli
    Claudio Gallicchio
    Vincenzo Lomonaco
    Salvatore Petroni
    Year: 2021
    Towards Functional Safety Compliance of Recurrent Neural Networks
    CAIP
    EAI
    DOI: 10.4108/eai.20-11-2021.2314139
Davide Bacciu1, Antonio Carta1, Daniele Di Sarli1, Claudio Gallicchio1, Vincenzo Lomonaco1,*, Salvatore Petroni2
  • 1: Department of Computer Science, University of Pisa, Pisa, Italy
  • 2: Legal and Compliance Department, Marelli Europe S.p.A., Turin, Italy
*Contact email: vincenzo.lomonaco@unipi.it

Abstract

Deploying Autonomous Driving systems requires facing some novel challenges for the Automotive industry. One of the most critical aspects that can severely compromise their deployment is Functional Safety. The ISO 26262 standard provides guidelines to ensure Functional Safety of road vehicles. However, this standard is not suitable to develop Artificial Intelligence based systems such as systems based on Recurrent Neural Networks (RNNs). To address this issue, in this paper we propose a new methodology, composed of three steps. The first step is the robustness evaluation of the RNN against inputs perturbations. Then, a proper set of safety measures must be defined according to the model’s robustness, where less robust models will require stronger mitigation. Finally, the functionality of the entire system must be extensively tested according to Safety Of The Intended Functionality (SOTIF) guidelines, providing quantitative results about the occurrence of unsafe scenarios, and by evaluating appropriate Safety Performance Indicators.