New patent application published for: SYSTEMS AND METHODS FOR PARSING LOG FILES USING CLASSIFICATION AND A PLURALITY OF NEURAL NETWORKS

SYSTEMS AND METHODS FOR PARSING LOG FILES USING CLASSIFICATION AND A PLURALITY OF NEURAL NETWORKS

Claims

1. A system, comprising: at least one processor; and at least one non-transitory memory storing instructions that, when executed by the at least one processor, cause the system to perform operations comprising: receiving unstructured data; selecting a neural network based on the unstructured data; resetting, based on selecting the neural network and while retaining a structure of the neural network, the neural network by resetting one or more of at least one weight of the neural network or at least one activation function of the neural network; and training, after resetting the neural network, the neural network to parse one or more of a type or a category of the unstructured data.
2. The system of claim 1, wherein selecting the neural network comprises: applying a classifier to the unstructured data to determine that a format of the unstructured data differs from a set of known formats, and selecting the neural network based on determining that the format of the unstructured data differs from a set of known formats.
3. The system of claim 1, wherein selecting the neural network comprises: applying a classifier to the unstructured data to determine that the unstructured data comprises a new category of unstructured data, identify, based on determining that that the unstructured data comprises the new category of unstructured data, an existing category of unstructured data, and selecting the neural network based on the existing category of unstructured data.
4. A method for training a neural network to parse unstructured data, the method comprising: applying a classifier to unstructured data; selecting a neural network based on applying the classifier to the unstructured data; resetting, based on selecting the neural network and while retaining a structure of the neural network, the neural network; and training, after resetting the neural network, the neural network to parse one or more of a type or a category of the unstructured data.
5. The method of claim 4, further comprising: receiving a log file that includes the unstructured data before applying the classifier to the unstructured data.
6. The method of claim 5, wherein the log file is generated by at least one application.
7. The method of claim 4, wherein applying the classifier to the unstructured data comprises applying the classifier to the unstructured data to determine that a format of the unstructured data differs from a set of known formats.
8. The method of claim 4, wherein applying the classifier to the unstructured data comprises applying the classifier to the unstructured data to identify an existing category similar to a new category of the unstructured data.
9. The method of claim 4, wherein applying the classifier to the unstructured data comprises applying the classifier to determine that the unstructured data comprises a new category of unstructured data by determining that an output of the classifier exceeds one or more thresholds of expected outputs associated with a plurality of neural networks.
10. The method of claim 4, wherein resetting the neural network comprises: resetting the neural network by resetting at least one weight to a default value.
11. The method of claim 4, wherein resetting the neural network comprises: resetting the neural network by resetting at least one activation function to a Heaviside step function.
12. The method of claim 4, further comprising: generating synthetic unstructured data to assist with training the neural network.
13. The method of claim 4, wherein training the neural network comprises transforming the unstructured data into structured data.
14. The method of claim 13, wherein the structured data comprises at least one of relational data, graphical data, or object-oriented data.
15. One or more non-transitory computer-readable media storing instructions that, when executed by at least one processor, cause the at least one processor to perform operations comprising: selecting a neural network based on unstructured data; resetting, based on selecting the neural network and while retaining a structure of the neural network, the neural network by resetting one or more of at least one weight of the neural network or at least one activation function of the neural network; and training, after resetting the neural network, the neural network to parse one or more of a type or a category of the unstructured data.
16. The one or more non-transitory computer-readable media of claim 15, wherein the operations further comprise: receiving a log file that includes the unstructured data before selecting the neural network.
17. The one or more non-transitory computer-readable media of claim 15, wherein selecting the neural network comprises applying a classifier to the unstructured data to determine that a format of the unstructured data differs from a set of known formats.
18. The one or more non-transitory computer-readable media of claim 15, wherein selecting the neural network comprises applying a classifier to the unstructured data to identify an existing category similar to a new category of the unstructured data.
19. The one or more non-transitory computer-readable media of claim 15, wherein resetting the neural network comprises: resetting the neural network by resetting the at least one weight to a default value.
20. The one or more non-transitory computer-readable media of claim 15, wherein training the neural network comprises transforming the unstructured data into structured data.