Method and apparatus to detect non-human users on computer systems
Methods and systems for malicious non-human user detection on computing devices are described. The method includes collecting, by a processing device, raw data corresponding to a user action, converting, by the processing device, the raw data to features, wherein the features represent characteristics of a human user or a malicious code acting as if it were the human user, and comparing, by the processing device, at least one of the features against a corresponding portion of a characteristic model to differentiate the human user from the malicious code acting as if it were the human user.
1. A method of improving a processing device having a client component and a server component, comprising:
collecting, by the client component, data corresponding to an action occurring at the processing device;
converting, by the client component, the data to one or more features, wherein the one or more features represent characteristics of a human user or a bot;
comparing, by the client component, at least one of the features against a learned model to determine, in a replay check, whether the at least one feature is indicative of a replay by the bot; and
upon a determination that the at least one feature is not indicative of a replay by the bot, further processing the one or more features by the server component against the learned model in a behavior check to determine whether the one or more features represent a behavior of the human user.
2. The method of claim 1 , further comprising processing the data to reduce an effect of noise, perturbation, or randomization.
3. The method of claim 1 , wherein the replay check comprises performing at least one of an exact match or a nearest neighbor match.
4. The method of claim 1 , further comprising hashing the data to perform the replay check.
5. The method of claim 1 , wherein collecting the data corresponding to the action further comprises receiving the data from an input device.
6. The method of claim 1 , further comprising rejecting the action in response to a detection that the action is from the bot and not the human user.
7. A non-transitory computer readable storage medium comprising instructions that, when executed by a processing device having a client component and a server component, cause the processing device to perform operations, comprising:
collecting, by the client component, data corresponding to an action occurring at the processing device;
converting, by the client component, the data to one or more features, wherein the one or more features represent characteristics of a human user or a bot;
comparing, by the client component, at least one of the features against a learned model to determine, in a replay check, whether the at least one feature is indicative of a replay by the bot; and
upon a determination that the at least one feature is not indicative of a replay by the malicious code, further processing the features by the server component against the learned model in a behavior check to determine whether the features represent a behavior of the human user.
8. The non-transitory computer readable storage medium comprising instructions of claim 7 , wherein the operations further comprise processing the data to reduce an effect of noise, perturbation, or randomization.
9. The non-transitory computer readable storage medium comprising instructions of claim 8 , wherein the replay check comprises performing at least one of an exact match check or a nearest neighbor match.
10. The non-transitory computer readable storage medium comprising instructions of claim 7 , wherein the operations further comprise hashing the data to perform the replay check.
11. The non-transitory computer readable storage medium comprising instructions of claim 7 , wherein collecting the data corresponding to the action further comprises receiving the data from an input device.
12. The non-transitory computer readable storage medium comprising instructions of claim 7 , wherein the operations further comprise rejecting the action in response to a detection that the action is from the bot and not the human user.
13. A computing system, comprising:
a data storage device; and
a processing device having a client component and a server component, coupled to the data storage device, to:
collect, by the client component, data corresponding to an action occurring at the processing device;
convert, by the client component, the data to one or more features, wherein the one or more features represent characteristics of a human user or a bot;
compare, by the client component, at least one of the features against a learned model to determine, in a replay check, whether the at least one feature is indicative of a replay by the bot; and
upon a determination that the at least one feature is not indicative of a replay by the bot, further process the features by the server component against the learned model in a behavior check to determine whether the features represent a behavior of the human user.
14. The system of claim 13 , wherein the processing device is further configured to process the data to reduce an effect of noise, perturbation, or randomization of the at least one of the features against the learned model.
15. The system of claim 13 , wherein the replay check comprises performing at least one of an exact match check or a nearest neighbor match.
16. The system of claim 13 , wherein the processing device is further configured to receive the data from an input device.
17. The system of claim 13 , wherein the processing device is further operative to reject the action in response to a detection that the action is from the bot and not the human user.