Mitigating Risk in Neural Network Classifiers

TY - GEN

T1 - Mitigating Risk in Neural Network Classifiers

AU - Alpizar Santana, Misael

AU - Calinescu, Radu

AU - Paterson, Colin

N1 - This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s self-archiving policy. Further copying may not be permitted; contact the publisher for details

PY - 2022/8/30

Y1 - 2022/8/30

N2 - Deep Neural Network (DNN) classifiers perform remarkably well on many problems that require skills which are natural and intuitive to humans. These classifiers have been used in safety-critical systems including autonomous vehicles. For such systems to be trusted it is necessary to demonstrate that the risk factors associated with neural network classification have been appropriately considered and sufficient risk mitigation has been employed. Traditional DNNs fail to explicitly consider risk during their training and verification stages, meaning that unsafe failure modes are permitted and under-reported.To address this limitation, our short paper introduces a work-in-progress approach that (i) allows the risk of misclassification between classes to be quantified, (ii) guides the training of DNN classifiers towards mitigating the risks that require treatment, and (iii) synthesises risk-aware ensembles with the aid of multi-objective genetic algorithms that seek to optimise DNN performance metrics while also mitigating risks.We show the effectiveness of our approach by using it to synthesise risk-aware neural network ensembles for the CIFAR-10 dataset.

AB - Deep Neural Network (DNN) classifiers perform remarkably well on many problems that require skills which are natural and intuitive to humans. These classifiers have been used in safety-critical systems including autonomous vehicles. For such systems to be trusted it is necessary to demonstrate that the risk factors associated with neural network classification have been appropriately considered and sufficient risk mitigation has been employed. Traditional DNNs fail to explicitly consider risk during their training and verification stages, meaning that unsafe failure modes are permitted and under-reported.To address this limitation, our short paper introduces a work-in-progress approach that (i) allows the risk of misclassification between classes to be quantified, (ii) guides the training of DNN classifiers towards mitigating the risks that require treatment, and (iii) synthesises risk-aware ensembles with the aid of multi-objective genetic algorithms that seek to optimise DNN performance metrics while also mitigating risks.We show the effectiveness of our approach by using it to synthesise risk-aware neural network ensembles for the CIFAR-10 dataset.

KW - deep neural network, risk, risk mitigation

M3 - Conference contribution

BT - 48th Euromicro Conference Series on Software Engineering and Advanced Applications (SEAA)

PB - IEEE

ER -