Systems and methods for artificial model building techniques
Embodiments disclosed describe a security awareness system may adaptively learn the best design of a simulated phishing campaign to get a user to perform the requested actions, such as clicking a hyperlink or opening a file. In some implementations, the system may adapt an ongoing campaign based on user's responses to messages in the campaign, along with the system's learned awareness. The learning process implemented by the security awareness system can be trained by observing the behavior of other users in the same company, other users in the same industry, other users that share similar attributes, all other users of the system, or users that have user attributes that match criteria set by the system, or that match attributes of a subset of other users in the system.
1. A method for establishing a model for communicating via simulated phishing campaigns, the method comprising:
(a) establishing, via one or more workers, a plurality of question and answer pairs to train a model for communicating via simulated phishing campaigns;
(b) training, by model trainer logic executing on a computing device, a neural network with the plurality of question and answer pairs, the model training adjusting settings of the neural network responsive to processing the plurality of question and answer pairs;
(c) establishing the model, by the model trainer logic responsive to training the neural network, the model to comprising a predetermined persona for simulated phishing communications and values corresponding to the adjusted settings of the neural network; and
(d) storing the model to be used by a campaign controller logic for communicating simulated phishing communications to one or more computing devices of one or more users.
2. The method of claim 1 , wherein (a) further comprises identifying, by the one or more workers, one or more of the questions or answers of the plurality of question and answer pairs from communications between the campaign controller logic and one or more users during execution of a simulated phishing campaign.
3. The method of claim 1 , wherein (a) further comprising creating, via the one or more workers, one or more of the questions or answers of the plurality of question and answer pairs.
4. The method of claim 3 , further comprising validating that answers to the questions of the plurality of question and answer pairs were established in accordance with the predetermined persona.
5. The method of claim 1 , wherein (a) further comprises establishing one or more of the questions or answers of the plurality of question and answer pairs in accordance with the predetermined persona.
6. The method of claim 1 , wherein (c) further comprises validating, via the one or more workers, output of the model responsive to one or more inputs to the model.
7. The method of claim 1 , wherein (a) further comprises validating, via the one or more workers, that one or more of the questions or answers of the plurality of question and answer pairs meet one or more predetermined criteria.
8. The method of claim 7 , wherein the one or more predetermined criteria comprises one or more of the following: a level of quality of a plurality of levels of quality, proper use of grammar and spelling errors.
9. The method of claim 1 , further comprising generating, responsive to training the neural network, a metagraph and one or more inputs to pass to a list of operations to execute the metagraph by the campaign controller logic.
10. The method of claim 9 , wherein (c) further comprises establishing the model, the model comprising the metagraph.
11. The method of claim 1 , further comprising adjusting settings of neurons of the neural network responsive to processing the plurality of question and answer pairs.
12. The method of claim 11 , wherein (c) further comprises establishing the model, the model comprising a matrix of values corresponding to the adjusted settings of neurons of the neural network.
13. The method of claim 1 , wherein (b) further comprises adjusting the model via a tuning process.
14. The method of claim 1 , wherein the model further comprises the predetermined persona including one of the following: an assistant, a travel agent, a tech support representative, a credit card company representative and a financial institution representative.
15. The method of claim 1 , wherein the model further comprises the predetermined persona corresponding to one of an industry, a demographic or an organizational level in a company.
16. A system for establishing a model for communicating via simulated phishing campaigns, the system comprising:
one or more computing devices comprising one or more hardware processors;
model trainer logic configured to execute on the one or more computing devices and to train a neural network with a plurality of question and answer pairs established via one or more workers and responsive to processing the plurality of question and answer pairs, adjusting settings of the neural network; and
responsive to training the neural network, establish, by the model training logic, a model comprising a predetermined persona for simulated phishing communication and values corresponding to the adjusted settings of the neural network; and
a database storage configured to store the model to be used by a campaign controller logic for communicating simulated phishing communications to one or more computing devices of one or more users.
17. The system of claim 16 , wherein one or more of the questions or answers of the plurality of questions and answer pairs are identified by one or more workers from communications between the campaign controller logic and one or more users during execution of a simulated phishing campaign.
18. The system of claim 16 , wherein one or more questions or answers of the plurality of question and answer pairs are created by the one or more workers.
19. The system of claim 16 , wherein one or more questions or answers of the plurality of question and answer pairs are established by the one or more workers in accordance with the predetermined persona.
20. The system of claim 19 , wherein answers to the questions of the plurality of question and answer pairs are validated by the one or more workers to be established in accordance with the predetermined persona.
21. The system of claim 16 , wherein output of the model is validated by the one or more workers responsive to one or more inputs to the model.
22. The system of claim 16 , wherein the one or more of the questions or answers of the plurality of question and answer pairs are validated by the one or more workers to meet one or more predetermined criteria.
23. The system of claim 22 , wherein the one or more predetermined criteria comprises one or more of the following: a level of quality of a plurality of levels of quality, proper use of grammar and spelling errors.
24. The system of claim 16 , wherein the model trainer logic is further configured to generate, responsive to training the neural network, a metagraph and one or more inputs to pass to a list of operations to execute of the metagraph by the campaign controller logic.
25. The system of claim 24 , wherein the model trainer logic is further configured to establish the model, the model comprising the metagraph.
26. The system of claim 16 , wherein the model trainer logic is further configured to adjust settings of neurons of the neural network responsive to processing the plurality of question and answer pairs.
27. The system of claim 16 , wherein the model trainer logic is further configured to establish the model, the model comprising a matrix of values corresponding to the adjusted settings of neurons of the neural network.
28. The system of claim 16 , wherein the model trainer logic is further configured to adjust the model via a tuning process.
29. The system of claim 16 , wherein the model further comprises the predetermined persona including one of the following: an assistant, a travel agent, a tech support representative, a credit card company representative and a financial institution representative.
30. The system of claim 16 , wherein the model further comprises the predetermined persona corresponding to one of an industry, a demographic or an organizational level in a company.