Machine learning correlator to infer network properties
Approaches provide for securing an electronic environment. A threat analysis service can obtain data for devices, users, and threats from disparate sources and can correlate users to devices and threats to build an understanding of an electronic environment's operational, organizational, and security concerns in order to provide customized security strategies and remediations. Additionally, the threat analysis service can develop a model of an electronic environment's behavior by monitoring and analyzing various the data from the data sources. The model can be updated such that the threat analysis service can tailor its orchestration to complement existing operational processes.
1. A computer-implemented method, comprising:
maintaining a plurality of data sets from a plurality of data sources, wherein individual ones of the data sets contain data about a plurality of assets, a plurality of users, or a plurality of threats associated with a network of computing assets;
augmenting the data in the data sets by classifying at least an asset in a first data set of the plurality of data sets, a user in a second data set of the plurality of data sets, and a threat in a third data set of the plurality of data sets;
receiving a query, wherein
the query is associated with a subject, and
the subject is an asset classification, a user classification, or a security threat classification;
selecting, based on the subject of the query, a machine learning correlator to infer a property of the network, wherein the machine learning correlator is selected from a set of correlators including (a) an asset-to-user correlator trained to identify correlations of data between one or more assets of the plurality of assets and one or more users of the plurality of users, (b) an asset-to-threat correlator trained to identify correlations of data between the one or more assets and one or more threats of the plurality of threats, and (c) user-to-threat correlator trained to identify correlations of data between the one or more users and the one or more threats;
generating a response to the query based on the property of the network inferred by the machine learning correlator selected; and
performing a security action affecting the subject, and the one or more assets, the one or more users, or the one or more threats, based on the inferring of the property of the network.
2. The computer-implemented method of claim 1 , further comprising:
classifying the asset in the first data set as a physical asset or a role asset using a machine learning asset classifier.
3. The computer-implemented method of claim 1 , further comprising:
classifying the user in the second data set as an employee type, a group type, or a role type using a machine learning user classifier.
4. The computer-implemented method of claim 1 , further comprising:
classifying the threat in the third data set as a target threat, an actor threat, or an activity threat using a machine learning threat classifier.
5. The computer-implemented method of claim 1 , further comprising:
correlating the asset, the user, and the threat among the first data set, the second data set, and the third data set.
6. The computer-implemented method of claim 1 , further comprising:
identifying the one or more users, the one or more threats, or the one or more assets associated with the subject.
7. A non-transitory computer readable storage medium comprising program instructions executable to:
maintain a plurality of data sets from a plurality of data sources, wherein individual ones of the data sets contain data about a plurality of assets, a plurality of users, or a plurality of threats associated with a network of computing assets;
augment the data in the data sets with classifications of at least an asset in a first data set of the plurality of data sets, a user in a second data set of the plurality of data sets, and a threat in a third data set of the plurality of data sets;
receive a query, wherein
the query is associated with a subject, and
the subject is an asset classification, a user classification, or a security threat classification;
select, based on the subject of the query, a machine learning correlator to infer a property of the network, wherein the machine learning correlator is selected from a set of correlators including (a) an asset-to-user correlator trained to identify correlations of data between one or more assets of the plurality of assets and one or more users of the plurality of users, (b) an asset-to-threat correlator trained to identify correlations of data between the one or more assets and one or more threats of the plurality of threats, and (c) user-to-threat correlator trained to identify correlations of data between the one or more users and the one or more threats;
generate a response to the query based on the property of the network inferred by the machine learning correlator selected; and
perform a security action affecting the subject, and the one or more assets, the one or more users, or the one or more threats, based on the inferring of the property of the network.
8. The non-transitory computer readable storage medium of claim 7 , wherein the program instructions are executable to:
classify the asset in the first data set as a physical asset or a role asset using a machine learning asset classifier.
9. The non-transitory computer readable storage medium of claim 7 , wherein the program instructions are executable to:
classify the user in the second data set as an employee type, a group type, or a role type using a machine learning user classifier.
10. The non-transitory computer readable storage medium of claim 7 , wherein the program instructions are executable to:
classify the threat in the third data set as a target threat, an actor threat, or an activity threat using a machine learning threat classifier.
11. The non-transitory computer readable storage medium of claim 7 , wherein the program instructions are executable to:
correlate the asset, the user, and the threat among the first data set, the second data set, and the third data set.
12. The non-transitory computer readable storage medium of claim 7 , wherein the program instructions are executable to:
identify the one or more users, the one or more threats, or the one or more assets associated with the subject.
13. A system comprising:
one or more processors; and
a memory coupled to the one or more processors, wherein the memory stores program instructions executable by the one or more processors to:
maintain a plurality of data sets from a plurality of data sources, wherein individual ones of the data sets contain data about a plurality of assets, a plurality of users, or a plurality of threats associated with a network of computing assets;
augment the data in the data sets with classifications of at least an asset in a first data set of the plurality of data sets, a user in a second data set of the plurality of data sets, and a threat in a third data set of the plurality of data sets;
receive a query, wherein
the query is associated with a subject, and
the subject is an asset classification, a user classification, or a security threat classification;
select, based on the subject of the query, a machine learning correlator to infer a property of the network, wherein the machine learning correlator is selected from a set of correlators including (a) an asset-to-user correlator trained to identify correlations of data between one or more assets of the plurality of assets and one or more users of the plurality of users, (b) an asset-to-threat correlator trained to identify correlations of data between the one or more assets and one or more threats of the plurality of threats, and (c) user-to-threat correlator trained to identify correlations of data between the one or more users and the one or more threats;
generate a response to the query based on the property of the network inferred by the machine learning correlator selected; and
perform a security action affecting the subject, and the one or more assets, the one or more users, or the one or more threats, based on the inferring of the property of the network.
14. The system of claim 13 , wherein the program instructions are executable to:
classify the asset in the first data set as a physical asset or a role asset using a machine learning asset classifier.
15. The system of claim 13 , wherein the program instructions are executable to:
classify the user in the second data set as an employee type, a group type, or a role type using a machine learning user classifier.
16. The system of claim 13 , wherein the program instructions are executable to:
classify the threat in the third data set as a target threat, an actor threat, or an activity threat using a machine learning threat classifier.
17. The system of claim 13 , wherein the program instructions are executable to:
correlate the asset, the user, and the threat among the first data set, the second data set, and the third data set.
18. The system of claim 13 , wherein the program instructions are executable to:
identify the one or more users, the one or more threats, or the one or more assets associated with the subject.