SYSTEM AND METHOD FOR USE OF GRAPH NEURAL NETWORKS IN IDENTITY MANAGEMENT ARTIFICIAL INTELLIGENCE SYSTEMS
Systems and methods for embodiments of a graph based artificial intelligence systems for identity management are disclosed. Embodiments of the identity management systems disclosed herein may utilize graph neural networks for the implementation of identity management components. An identity management system may create an identity graph from identity management data. An embedding for the identity graph representing the identity management data for the enterprise may be generated and that embedding used as the basis for training a graph neural network for an identity management component. Once a graph neural network is trained for a particular identity management component, the identity management component can apply the associated trained graph neural network during operation of that component in the identity management system.
1 . An identity management system, comprising:
a data store;
a processor;
a non-transitory, computer-readable storage medium, including computer instructions for:
obtaining identity management data from one or more source systems in a distributed enterprise computing environment of an enterprise, the identity management data comprising data on a set of identities, a set of entitlements, or a set of roles, wherein the set of identities, set of entitlements or set of roles are utilized in identity management in the distributed enterprise computing environment;
generating a first identity graph from the identity management data at a first time;
training a first graph neural network for a first identity management component; and
adapting the first identity management component to use the first graph neural network such that the first identity management component is adapted to generate a first identity management signal using the first graph neural network.
2 . The system of claim 1 , wherein training the first graph neural network comprises:
generating a first embedding from the first graph neural network; and
training the first graph neural network based on the first embedding and a first loss function associated with the first identity management component.
3 . The system of claim 2 , wherein the first identity management signal is associated with clustering of the identity graph.
4 . The system of claim 3 , wherein the first loss function is a spectral loss version of modularity.
5 . The system of claim 2 , comprising:
training a second graph neural network based on the embedding and a second loss function associated with a second identity management component; and
adapting the second identity management component to generate a second identity management signal using the second graph neural network.
6 . The system of claim 5 , comprising:
updating the identity management data;
generating a second identity graph from the updated identity management data at a second time;
training a second graph neural network for the first identity management component; and
adapting the first identity management component to use the second graph neural network such that the first identity management component is adapted to generate the first identity management signal using the second graph neural network.
7 . The system of claim 6 , wherein training the second graph neural network comprises:
generating a second embedding from the second graph neural network; and
training the second graph neural network based on the second embedding and the first loss function associated with the first identity management component.
8 . A method for identity management, comprising:
obtaining identity management data from one or more source systems in a distributed enterprise computing environment of an enterprise, the identity management data comprising data on a set of identities, a set of entitlements, or a set of roles, wherein the set of identities, set of entitlements or set of roles are utilized in identity management in the distributed enterprise computing environment;
generating a first identity graph from the identity management data at a first time;
training a first graph neural network for a first identity management component; and
adapting the first identity management component to use the first graph neural network such that the first identity management component is adapted to generate a first identity management signal using the first graph neural network.
9 . The method of claim 8 , wherein training the first graph neural network comprises:
generating a first embedding from the first graph neural network; and
training the first graph neural network based on the first embedding and a first loss function associated with the first identity management component.
10 . The method of claim 9 , wherein the first identity management signal is associated with clustering of the identity graph.
11 . The method of claim 10 , wherein the first loss function is a spectral loss version of modularity.
12 . The method of claim 9 , further comprising:
training a second graph neural network based on the embedding and a second loss function associated with a second identity management component; and
adapting the second identity management component to generate a second identity management signal using the second graph neural network.
13 . The method of claim 12 , further comprising:
updating the identity management data;
generating a second identity graph from the updated identity management data at a second time;
training a second graph neural network for the first identity management component; and
adapting the first identity management component to use the second graph neural network such that the first identity management component is adapted to generate the first identity management signal using the second graph neural network.
14 . The method of claim 13 , wherein training the second graph neural network comprises:
generating a second embedding from the second graph neural network; and
training the second graph neural network based on the second embedding and the first loss function associated with the first identity management component.
15 . A non-transitory computer readable medium, comprising instructions for:
obtaining identity management data from one or more source systems in a distributed enterprise computing environment of an enterprise, the identity management data comprising data on a set of identities, a set of entitlements, or a set of roles, wherein the set of identities, set of entitlements or set of roles are utilized in identity management in the distributed enterprise computing environment;
generating a first identity graph from the identity management data at a first time;
training a first graph neural network for a first identity management component; and
adapting the first identity management component to use the first graph neural network such that the first identity management component is adapted to generate a first identity management signal using the first graph neural network.
16 . The non-transitory computer readable medium of claim 15 , wherein training the first graph neural network comprises:
generating a first embedding from the first graph neural network; and
training the first graph neural network based on the first embedding and a first loss function associated with the first identity management component.
17 . The non-transitory computer readable medium of claim 16 , wherein the first identity management signal is associated with clustering of the identity graph.
18 . The non-transitory computer readable medium of claim 17 , wherein the first loss function is a spectral loss version of modularity.
19 . The non-transitory computer readable medium of claim 16 , further comprising instructions for:
training a second graph neural network based on the embedding and a second loss function associated with a second identity management component; and
adapting the second identity management component to generate a second identity management signal using the second graph neural network.
20 . The non-transitory computer readable medium of claim 19 , further comprising instructions for:
updating the identity management data;
generating a second identity graph from the updated identity management data at a second time;
training a second graph neural network for the first identity management component; and
adapting the first identity management component to use the second graph neural network such that the first identity management component is adapted to generate the first identity management signal using the second graph neural network.
21 . The non-transitory computer readable medium of claim 20 , wherein training the second graph neural network comprises:
generating a second embedding from the second graph neural network; and
training the second graph neural network based on the second embedding and the first loss function associated with the first identity management component.