Improving Robustness of Deep Neural Networks for Aerial Navigation by Incorporating Input Uncertainty.
Fabio ArnezHuáscar EspinozaAnsgar RadermacherFrançois TerrierPublished in: SAFECOMP Workshops (2021)
Keyphrases
- neural network
- pattern recognition
- artificial neural networks
- fuzzy logic
- incomplete information
- desired output
- computational efficiency
- multi layer
- neural nets
- uncertain data
- vehicle tracking
- inherent uncertainty
- obstacle avoidance
- expected utility
- multilayer perceptron
- hidden units
- user input
- satellite images
- feed forward
- neural network model
- sufficient conditions
- decision makers
- probability distribution
- genetic algorithm