| CPC H04L 63/0407 (2013.01) [H04L 41/16 (2013.01)] | 20 Claims |
|
1. A system for monitoring data networks featuring data traffic using probabilistic graphical models, the system comprising:
one or more processors; and
one or more non-transitory, computer-readable media comprising instructions that, when executed by the one or more processors, cause operations comprising:
receiving, via a user interface, a first user request to analyze network data traffic over a first cloud computing network with a first model, wherein the first model directed to a deep learning network with a plurality of unknown characteristics of the first model, wherein the plurality of unknown characteristics is used to process inputs to the first model to generate outputs, and
wherein the plurality of unknown characteristics include one or more of features, output variables or targets, latent variables, control variables for decision models, environment variables, model parameters, attributes and properties affecting the first model's behavior, capabilities, performance or weight of a node, or edge, of the deep learning network being used to produce a result;
receiving, via the user interface, a predefined data threshold, wherein the predefined data threshold comprises a requirement for a threshold level of data security when processing user data through the first model to determine whether the first model corresponds to compliance requirements;
generating a second model corresponding to the first model, wherein the second model is an observer model that does not perform a feature importance analysis in which an importance score is attributed to the feature or the variable in the second model by performing one or more permutations on a given feature, wherein the second model comprises a probabilistic graphical model corresponding to the first model,
wherein the probabilistic graphical model is trained to determine probabilities for graphical characteristics corresponding to the plurality of unknown characteristics, wherein the probabilities corresponding to the graphical characteristics correspond to a probability that the node, edge, or weight of the first model was used to generate the result when processing the user data through the first model, and wherein the probabilistic graphical model is trained to determine probabilities for graphical characteristics corresponding to the plurality of unknown characteristics by:
receiving training data, wherein the training data comprises permutation importance values for features in the first model;
aggregating the permutation importance values to generate an aggregated set; and
training the second model based on the aggregated set;
determining a first graphical characteristic of the graphical characteristics corresponding to the predefined data threshold;
comparing a first probability to a threshold probability to determine whether the first model corresponds to the predefined data threshold; and
generating for display, on a user interface, a recommendation based on comparing the first probability to the threshold probability, wherein the threshold probability are used for classification decision-making tasks and make binary predictions.
|