Researchers in this field have now proposed a hybrid architecture along with a biologically-constrained neural network-VOneNets for the creation of computer vision models making them more robust for withstanding the adversarial attacks.
The convolutional neural networks had been dominating the field of object recognition this CNN had the flow of being easily deceived merely with the creation of a small perturbation which is also termed as adversarial attacks. This will further lead to the failure of the computer vision models and make it susceptible to cyberattacks. CNN’s vulnerability to the attacks of image perturbations had become a pressing concern for the machine learning community while researchers and scientists are working towards building computer vision models that generalise images like humans.
For addressing the vulnerability, researchers from MIT, Harvard University and MIT-IBM Watson AI Lab have proposed VOneNets- the new class of hybrid CNN vision models-in a recent paper. According to the researchers, this novel architecture leverages “biologically-constrained neural networks along with deep learning techniques” to create more model robustness against white-box adversarial attacks.
Explaining the process, the researchers stated that VOneNets is going to be a new class of CNN’s, that “contain a biologically-constrained neural network that simulates the primary visual cortex of primates.” The researchers noted that VOneNet replaces the first few layers with the VOneBlock, known as V1 front-end or primary visual cortex of primates.
The susceptibility of convolutional neural networks to small image perturbations suggests that these neural networks rely on visual features that are not used by the primates. Inspired by these studies that showcased the high correlation between the explained variance of the brain’s primary visual cortex, V1 front-end, and CNN’s robustness to white-box attacks, the researcher’s team developed VOneNet architecture.
V1 front-end, with its “fixed-weight, simple and complex cell nonlinearities, and neuronal stochasticity,” is what characterises VOneNet, with its “fixed-weight, simple and complex cell nonlinearities and neuronal stochasticity,” is what characterises VOneNet. This can also be adapted to different CNN-based architectures such as ResNet, CORnet-S and AlexNet.
To conclude the following whilst experimenting with VoOneNet showcased that neuroscience still has a lot of untapped potentials to solve critical AI problems. The model requires “less training to achieve human intelligence,’ which can advance more neuroscience-inspired machine learning algorithms.