Systems, methods, and media for generating sanitized data, sanitizing anomaly detection models, and/or generating sanitized anomaly detection models
Systems, methods, and media for generating sanitized data, sanitizing anomaly detection models, and generating anomaly detection models are provided. In some embodiments, methods for sanitizing anomaly detection models are provided. The methods including: receiving at least one abnormal anomaly detection model from at least one remote location; comparing at least one of the at least one abnormal anomaly detection model to a local normal detection model to produce a common set of features common to both the at least one abnormal anomaly detection model and the local normal detection model; and generating a sanitized normal anomaly detection model by removing the common set of features from the local normal detection model.
1. A method for sanitizing anomaly detection models, the method comprising:
receiving at least one abnormal anomaly detection model from at least one remote device, wherein the at least one abnormal anomaly detection model includes at least one model of data items that have been selected based on an anomaly score, and wherein the anomaly score for each of the data items is calculated based on a weight being applied to a label corresponding to a data item;
comparing at least one of the at least one abnormal anomaly detection model to a local normal detection model to determine content contained in both the at least one abnormal anomaly detection model and the local normal detection model; and
generating a sanitized normal anomaly detection model by removing the content contained in both the at least one abnormal anomaly detection model and the local normal detection model from the local normal detection model.
2. The method of claim 1 , further comprising generating a local abnormal anomaly detection model based on the content contained in both the at least one abnormal anomaly detection model and the local normal detection model.
3. The method of claim 1 , further comprising updating the at least one abnormal anomaly detection model with the content contained in both the at least one abnormal anomaly detection model and the local normal detection model.
4. The method of claim 1 , wherein the at least one abnormal anomaly detection model and the local normal detection model are shared between remote devices.
5. The method of claim 1 , wherein the at least one abnormal anomaly detection model and the local normal detection model are combined to form a detection model, wherein a first portion of the detection model is identified as abnormal based on the at least one abnormal anomaly detection model, and wherein a second portion of the detection model is identified as normal based on the local normal detection model.
6. The method of claim 1 , further comprising testing an input dataset using the sanitized normal anomaly detection model to determine if the input dataset includes anomalies.
7. The method of claim 1 , further comprising:
testing at least one data item of a training dataset of data items against the at least one abnormal anomaly detection model to provide an indication that the at least one data item is abnormal; and
creating a sanitized training dataset of data items based on the indication that the at least one data item is abnormal.
8. The method of claim 7 , further comprising generating a sanitized anomaly detection model based on the sanitized training dataset.
9. The method of claim 7 , further comprising creating an abnormal training dataset of data items based on the indication that the at least one data item is abnormal.
10. The method of claim 9 , further comprising generating an abnormal anomaly detection model based on the anomalous training dataset.
11. A system for sanitizing anomaly detection models, the system comprising:
an interface in communication with a network;
a memory; and
a processor in communication with the memory and the interface;
wherein the processor:
receives at least one abnormal anomaly detection model from at least one remote device, wherein the at least one abnormal anomaly detection model includes at least one model of data items that have been selected based on an anomaly score, and wherein the anomaly score for each of the data items is calculated based on a weight being applied to a label corresponding to a data item;
compares at least one of the at least one abnormal anomaly detection model to a local normal detection model to determine content contained in both the at least one abnormal anomaly detection model and the local normal detection model; and
generates a sanitized normal anomaly detection model by removing the content contained in both the at least one abnormal anomaly detection model and the local normal detection model from the local normal detection model.
12. The system of claim 11 , wherein the processor is further configured to generate a local abnormal anomaly detection model based on the content contained in both the at least one abnormal anomaly detection model and the local normal detection model.
13. The system of claim 11 , wherein the processor is further configured to update the at least one abnormal anomaly detection model with the content contained in both the at least one abnormal anomaly detection model and the local normal detection model.
14. The system of claim 11 , wherein the at least one abnormal anomaly detection model and the local normal detection model are shared between remote devices.
15. The system of claim 11 , wherein the at least one abnormal anomaly detection model and the local normal detection model are combined to form a detection model, wherein a first portion of the detection model is identified as abnormal based on the at least one abnormal anomaly detection model, and wherein a second portion of the detection model is identified as normal based on the local normal detection model.
16. The system of claim 11 , wherein the processor is further configured to test an input dataset using the sanitized normal anomaly detection model to determine if the input dataset includes anomalies.
17. The system of claim 11 , wherein the processor is further configured to:
test at least one data item of a training dataset of data items against the at least one abnormal anomaly detection model to provide an indication that the at least one data item is abnormal; and
create a sanitized training dataset of data items based on the indication that the at least one data item is abnormal.
18. The system of claim 17 , wherein the processor is further configured to generate a sanitized anomaly detection model based on the sanitized training dataset.
19. The system of claim 17 , wherein the processor is further configured to create an abnormal training dataset of data items based on the indication that the at least one data item is abnormal.
20. The system of claim 19 , wherein the processor is further configured to generate an abnormal anomaly detection model based on the anomalous training dataset.
21. A non-transitory computer-readable medium containing computer-executable instructions that, when executed by a processor, cause the processor to perform a method for sanitizing anomaly detection models, the method comprising:
receiving at least one abnormal anomaly detection model from at least one remote device, wherein the at least one abnormal anomaly detection model includes at least one model of data items that have been selected based on an anomaly score, and wherein the anomaly score for each of the data items is calculated based on a weight being applied to a label corresponding to a data item;
comparing at least one of the at least one abnormal anomaly detection model to a local normal detection model to determine content contained in both the at least one abnormal anomaly detection model and the local normal detection model; and
generating a sanitized normal anomaly detection model by removing the content contained in both the at least one abnormal anomaly detection model and the local normal detection model from the local normal detection model.
22. A method for generating anomaly detection models, the method comprising:
receiving at least one abnormal anomaly detection model from at least one remote device, wherein the at least one abnormal anomaly detection model includes at least one model of data items that have been selected based on an anomaly score, and wherein the anomaly score for each of the data items is calculated based on a weight being applied to a label corresponding to a data item;
testing at least one data item of a training dataset of data items against the at least one abnormal anomaly detection model to provide an indication that the at least one data item is abnormal;
creating a sanitized training dataset of data items based on the indication that the at least one data item is abnormal;
generating a sanitized anomaly detection model based on the sanitized training dataset;
creating an abnormal training dataset of data items based on the indication that the at least one data item is abnormal; and
generating an abnormal anomaly detection model based on the anomalous training dataset.
23. A system for generating anomaly detection models, the system comprising:
a hardware processor that is configured to:
receive at least one abnormal anomaly detection model from at least one remote device, wherein the at least one abnormal anomaly detection model includes at least one model of data items that have been selected based on an anomaly score, and wherein the anomaly score for each of the data items is calculated based on a weight being applied to a label corresponding to a data item;
test at least one data item of a training dataset of data items against the at least one abnormal anomaly detection model to provide an indication that the at least one data item is abnormal;
create a sanitized training dataset of data items based on the indication that the at least one data item is abnormal;
generate a sanitized anomaly detection model based on the sanitized training dataset;
create an abnormal training dataset of data items based on the indication that the at least one data item is abnormal; and
generate an abnormal anomaly detection model based on the anomalous training dataset.
24. A non-transitory computer-readable medium containing computer executable instructions that, when executed by a processor, cause the processor to perform a method for generating anomaly detection models, the method comprising:
receiving at least one abnormal anomaly detection model from at least one remote device, wherein the at least one abnormal anomaly detection model includes at least one model of data items that have been selected based on an anomaly score, and wherein the anomaly score for each of the data items is calculated based on a weight being applied to a label corresponding to a data item;
testing at least one data item of a training dataset of data items against the at least one abnormal anomaly detection model to provide an indication that the at least one data item is abnormal;
creating a sanitized training dataset of data items based on the indication that the at least one data item is abnormal;
generating a sanitized anomaly detection model based on the sanitized training dataset;
creating an abnormal training dataset of data items based on the indication that the at least one data item is abnormal; and
generating an abnormal anomaly detection model based on the anomalous training dataset.