Abstract: Method and system for watermarking prediction outputs generated by a first AI model to enable detection of a target AI model that has been distilled from the prediction outputs. Includes receiving, at the first AI model, a set of input data samples from a requesting device; storing at least a subset of the input data samples to maintain a record of the input data samples; predicting, using the first AI model, a respective set of prediction outputs that each include a probability value, the AI model using a watermark function to insert a periodic watermark signal in the probability values of the prediction outputs; and outputting, from the first AI model, the prediction outputs including the periodic watermark signal.

Abstract: A machine learning model is learned using secure vertical federated learning by receiving, by a network machine learning model, from a plurality of private machine learning models, a set of private machine learning model outputs. The set of private machine learning model outputs is based on data owned exclusively by each of the plurality of private machine learning models. The set of private machine learning model machine learning outputs is aligned based on sample IDs of the data. The network machine learning model, a prediction, the prediction being the output of the network model based on the set of private machine learning model outputs. Transmitting, by the network model, the prediction, to one of the plurality of private machine learning models, the one of the plurality of private machine learning models comprising labels.

Abstract: Methods and systems for mining minority-class data samples are described. A minority-class mining service receives activations generated by an inner-layer of a client neural network that has been trained to perform a prediction task that involves classification. The minority-class mining service generates a recalibrated activation using a recalibration neural network, and generates an anomaly detector output using an anomaly detector. From the anomaly detector output, a minority-class score is computed for the data sample represented by a received activation. The computed minority-class score is compared against a minority-class threshold to identify a candidate minority-class data sample. The candidate minority-class data sample can then be labeled and added to the training dataset for the client neural network.

Abstract: Methods and systems for horizontal federated learning are described. A plurality of sets of local model parameters is obtained. Each set of local model parameters was learned at a respective client. For each given set of local model parameters, collaboration coefficients are computed, representing a similarity between the given set of local model parameters and each other set of local model parameters. Updating of the sets of local model parameters is performed, to obtain sets of updated local model parameters. Each given set of local model parameters is updated using a weighted aggregation of the other sets of local model parameters, where the weighted aggregation is computed using the collaboration coefficients. The sets of updated local model parameters are provided to each respective client.

Abstract: The present disclosure provides a method and system for evaluating a machine learning model using an evaluation dataset for the machine learning model. The evaluation dataset includes for each entity in a group of entities: (i) an ordered set of attribute values for the entity, each attribute value corresponding to a respective attribute in a set of attributes that is common for all of the entities in the group of entities, and (ii) an outcome prediction generated for the entity by the machine learning model based on the ordered set of attribute values for the entity, wherein the outcome prediction generated for each entity is either a first outcome or a second outcome. Based on the evaluation dataset, using an optimization process, respective importance values are computed for the attributes, the respective importance values indicating attributes that are most responsible for the machine learning model predicting a first outcome.

Abstract: Method and system for watermarking prediction outputs generated by a first AI model to enable detection of a target AI model that has been distilled from the prediction outputs. Includes receiving, at the first AI model, a set of input data samples from a requesting device; storing at least a subset of the input data samples to maintain a record of the input data samples; predicting, using the first AI model, a respective set of prediction outputs that each include a probability value, the AI model using a watermark function to insert a periodic watermark signal in the probability values of the prediction outputs; and outputting, from the first AI model, the prediction outputs including the periodic watermark signal.

Abstract: Methods, systems and media for deep neural network interpretation via rule extraction. The interpretation of the deep neural network is based on extracting one or more rules approximating classification behavior of the network. Rules are defined by identifying a set of hyperplanes through the data space that collectively define a convex polytope that separates a target class of input samples from input samples of different classes. Each rule corresponds to a set of decision boundaries between two different decision outcomes. Human-understandable representations of rules may be generated. One or more rules may be used to generate a classifier. The representations and interpretations exhibit faithfulness, robustness, and comprehensiveness relative to other known approaches.

Abstract: Methods and systems for mining minority-class data samples are described. A minority-class mining service receives activations generated by an inner-layer of a client neural network that has been trained to perform a prediction task that involves classification. The minority-class mining service generates a recalibrated activation using a recalibration neural network, and generates an anomaly detector output using an anomaly detector. From the anomaly detector output, a minority-class score is computed for the data sample represented by a received activation. The computed minority-class score is compared against a minority-class threshold to identify a candidate minority-class data sample. The candidate minority-class data sample can then be labeled and added to the training dataset for the client neural network.

Abstract: Methods, systems and media for deep neural network interpretation via rule extraction. The interpretation of the deep neural network is based on extracting one or more rules approximating classification behavior of the network. Rules are defined by identifying a set of hyperplanes through the data space that collectively define a convex polytope that separates a target class of input samples from input samples of different classes. Each rule corresponds to a set of decision boundaries between two different decision outcomes. Human-understandable representations of rules may be generated. One or more rules may be used to generate a classifier. The representations and interpretations exhibit faithfulness, robustness, and comprehensiveness relative to other known approaches.

Abstract: Methods and systems for horizontal federated learning are described. A plurality of sets of local model parameters is obtained. Each set of local model parameters was learned at a respective client. For each given set of local model parameters, collaboration coefficients are computed, representing a similarity between the given set of local model parameters and each other set of local model parameters. Updating of the sets of local model parameters is performed, to obtain sets of updated local model parameters. Each given set of local model parameters is updated using a weighted aggregation of the other sets of local model parameters, where the weighted aggregation is computed using the collaboration coefficients. The sets of updated local model parameters are provided to each respective client.

Abstract: A machine learning model is learned using secure vertical federated learning by receiving, by a network machine learning model, from a plurality of private machine learning models, a set of private machine learning model outputs. The set of private machine learning model outputs is based on data owned exclusively by each of the plurality of private machine learning models. The set of private machine learning model machine learning outputs is aligned based on sample IDs of the data. The network machine learning model, a prediction, the prediction being the output of the network model based on the set of private machine learning model outputs. Transmitting, by the network model, the prediction, to one of the plurality of private machine learning models, the one of the plurality of private machine learning models comprising labels.