Abstract: Techniques for training and using models to analyze multiple attributes of an authentication, and detect anomalous authentications that may include security threats. An authentication platform may use historical authentication data to train models to identify common attributes for authentications of users, and the training may be performed without the use of labels. For instance, models may be trained for each attribute of the historical authentications (e.g., geographic location, type of authentication method, type of device, time of day, etc.) to “learn” common behaviors of users across attributes of historical authentications. The models can then be applied to new authentications to determine, on an attribute-by-attribute level, whether or not new authentications are anomalous as compared to historical authentications by the user.

Abstract: Techniques for training and using models to analyze multiple attributes of an authentication, and detect anomalous authentications that may include security threats. An authentication platform may use historical authentication data to train models to identify common attributes for authentications of users, and the training may be performed without the use of labels. For instance, models may be trained for each attribute of the historical authentications (e.g., geographic location, type of authentication method, type of device, time of day, etc.) to “learn” common behaviors of users across attributes of historical authentications. The models can then be applied to new authentications to determine, on an attribute-by-attribute level, whether or not new authentications are anomalous as compared to historical authentications by the user.

Abstract: In one embodiment, a machine learning model evaluation system may define standardized, extensible class hierarchies for evaluating performance of a given machine learning model. The class hierarchies may include a plurality of target classes that formalize an expected output of the given machine learning model based on a given dataset, a plurality of output classes that formalize an actual output of the given machine learning model based on the given dataset, a plurality of metric classes that formalize a comparison of the expected output of the given machine learning model with the actual output of the given machine learning model, and a plurality of datasets. When a machine learning model is received for evaluation, the system may identify a target class, an output class, and a metric class that are applicable to the machine learning model. The system may also retrieve a dataset applicable to the machine learning model.

Abstract: In one embodiment, a machine learning model evaluation system may define standardized, extensible class hierarchies for evaluating performance of a given machine learning model. The class hierarchies may include a plurality of target classes that formalize an expected output of the given machine learning model based on a given dataset, a plurality of output classes that formalize an actual output of the given machine learning model based on the given dataset, a plurality of metric classes that formalize a comparison of the expected output of the given machine learning model with the actual output of the given machine learning model, and a plurality of datasets. When a machine learning model is received for evaluation, the system may identify a target class, an output class, and a metric class that are applicable to the machine learning model. The system may also retrieve a dataset applicable to the machine learning model.