Abstract: A system and method for machine learning architecture for prospective resource allocations are described. The method may include: receiving data records representing historical resource allocations from a user account associated with a first identifier to a resource account associated with a second identifier; deriving input features based on the data records; computing, using a trained neural network architecture, a predicted resource allocation amount and a predicted resource allocation date for the predicted resource allocation amount based on the derived input features; determining, using the trained neural network architecture, a first selection score associated with the predicted resource allocation amount and a second selection score associated with the predicted resource allocation date; and when the first or second selection score is above a minimum threshold, causing to display, at a display device, the associated resource allocation amount or date corresponding to the second identifier.

Abstract: A system and method for adversarial vulnerability testing of machine learning models is proposed that receives as an input, a representation of a non-differentiable machine learning model, transforms the input model into a smoothed model and conducts an adversarial search against the smoothed model to generate an output data value representative of a potential vulnerability to adversarial examples. Variant embodiments are also proposed, directed to noise injection, hyperparameter control, and exhaustive/sampling-based searches in an effort to balance computational efficiency and accuracy in practical implementation. Flagged vulnerabilities can be used to have models re-validated, re-trained, or removed from use due to an increased cybersecurity risk profile.

Abstract: Systems and methods for machine learning architecture for out-of-distribution data detection. The system may include a processor and a memory storing processor-executable instructions that may, when executed, configure the processor to: receive an input data set; generate an out-of-distribution prediction based on the input data set and an auto-encoder, the auto-encoder trained based on a pretext task including a transformation of one or more training data sets for reconstruction, the trained auto-encoder trained for reducing a reconstruction error to encode semantic meaning of the training data sets; and generate a signal for providing an indication of whether the input data set is an out-of-distribution data set.

Abstract: Systems and methods for diagnosing and testing fairness of machine learning models based on detecting individual violations of group definitions of fairness, via adversarial attacks that aim to perturb model inputs to generate individual violations. The systems and methods employ auxiliary machine learning models using a local surrogate for identifying group membership and assess fairness by measuring the transferability of attacks from this model. The systems and methods generate fairness indicator values indicative of discrimination risk due to the target predictions generated by the machine learning model, by comparing gradients of the machine learning model to gradients of an auxiliary machine learning model.

Abstract: An approach for increasing security of biometric templates is described. An improved system is adapted to split a full set of features or representations of a trained model into a first partial template and a second partial template, the second partial template being stored on a secure enclave accessible only through zero-knowledge proof based interfaces. During verification using the template, a new full set of features is received for comparison, and a model is loaded based on the available portions of the model. Comparison utilizing the second partial template requires the computation of zero-knowledge proofs as direct access to the underlying second partial template is prohibited by the secure enclave.