Charles River Analytics has secured funds from the Department of Defense's Strategic Capabilities Office to provide a neural network intended to protect sensitive information against white or black box attacks.
The company said Wednesday it will develop the Secure Private Neural Network for analysts to train deep neural networks with data protected via end-to-end encryption.
Adversarial exploitation of DNNs can result in a data breach or misclassification incident, the company noted.
“Cyber adversaries can monitor deep neural networks and learn their training and classification processes,” said Curt Wu, chief software engineer at Charles River and SPNN project manager.
“SPNN uses privacy-preserving encryption so deep neural networks can securely perform training and classification tasks,” Wu added.
The network will add to the firm's deep learning portfolio, which includes the Causal Models to Explain Learning framework. CAMEL is designed to support human-artificial intelligence dialogues.