Unstructured data identification and workflow execution using machine-learning techniques
The disclosed techniques are directed to identifying textual data instances depicted within images having an unstructured/undefined format. A machine-learning model may be trained to identify textual data instances within the image and corresponding data types for the textual data instances. The values and/or data types of the textual data instances may be compared to previously-stored data that is associated with a data provider. If the values and/or data types match the previously-stored data, the values corresponding to the textual data instances may be used to execute one or more processes. Executing a process may comprise transmitting one or more data messages that include one or more values of the textual data instances. The disclosed techniques may be executed as part of a monitoring process that obtains images over a time period, detects and validates the textual data instances depicted within those images, and executes one or more additional processes using values extracted from the images.
1 . A computer-implemented method, comprising:
receiving, by a computing device, an image having an unstructured format;
providing, by the computing device, the image as input to a machine-learning model that has been previously trained to identify a set of textual data instances within the image and a set of data types corresponding to the set of textual data instances, the machine-learning model being previously trained utilizing a machine-learning algorithm and a training data set comprising example images of a plurality of unstructured formats, each example image identifying corresponding textual data instances within a corresponding image of a respective unstructured format of the plurality of unstructured formats and corresponding data types of the corresponding textual data instances;
obtaining, by the computing device from the machine-learning model, output identifying the set of textual data instances within the image and the set of data types corresponding to the set of textual data instances;
determining, by the computing device, whether a textual data instance of the set of textual data instances is valid based at least in part on identifying a data type for the textual data instance and comparing a first value corresponding to the textual data instance to a second value of historical data associated with one or more users; and
based at least in part on determining that the textual data instance is valid, executing, by the computing device, operations that cause an automated process to be executed using the one or more values corresponding to the set of textual data instances.
2 . The computer-implemented method of claim 1 , wherein determining whether the textual data instance of the set of textual data instances is valid further comprises determining whether the first value corresponding to the textual data instance matches or falls within a difference threshold of the second value of the historical data associated with the one or more users.
3 . The computer-implemented method of claim 1 , further comprising obtaining, by the computing device, the historical data associated with the one or more users from a second textual data instance of the set of textual data instances based at least in part on identifying that the second textual data instance is associated with a second data type of the set of data types.
4 . The computer-implemented method of claim 1 , wherein the historical data associated with the one or more users comprises historical transactions of one or more accounts corresponding to the one or more users.
5 . The computer-implemented method of claim 1 , wherein the automated process is one of a plurality of automated processes, and wherein the computer-implemented method further comprises:
identifying, by the computing device, a second textual data instance of the set of textual data instances based on a second data type identified for the second textual data instance
identifying, by the computing device, an approved process type based at least in part on a value corresponding to the second textual data instance; and
selecting the automated process from the plurality of automated processes based at least in part on the approved process type that was identified.
6 . The computer-implemented method of claim 1 , wherein the image is associated with an online account, and wherein the online account and the first value corresponding to the textual data instance are utilized to execute the automated process.
7 . The computer-implemented method of claim 1 , wherein the image is one of a plurality of images obtained by the computing device over a time period, and wherein the computing device obtains the plurality of images based at least in part on receiving user input indicating a periodicity at which images are to be requested from a second computing device.
8 . A computing device, comprising:
one or more processors; and
one or more memories storing computer-executable instructions that, when executed by the one or more processors, cause the one or more processors to:
receive an image having an unstructured format;
provide the image as input to a machine-learning model that has been previously trained to identify a set of textual data instances within the image and a set of data types corresponding to the set of textual data instances, the machine-learning model being previously trained utilizing a machine-learning algorithm and a training data set comprising example images of a plurality of unstructured formats, each example identifying corresponding textual data instances within a corresponding image of a respective unstructured format of the plurality of unstructured formats and corresponding data types of the corresponding textual data instances;
obtain, from the machine-learning model, output identifying the set of textual data instances within the image and the set of data types corresponding to the set of textual data instances;
determine whether a textual data instance of the set of textual data instances is valid based at least in part on identifying a data type for the textual data instance and comparing a first value corresponding to the textual data instance to a second value of historical data associated with one or more users; and
based at least in part on determining that the textual data instance is valid, execute operations that cause an automated process to be executed using the one or more values corresponding to the set of textual data instances.
9 . The computing device of claim 8 , wherein executing the computer-executable instructions that determine whether the textual data instance of the set of textual data instances is valid further causes the one or more processors to determine whether the first value corresponding to the textual data instance matches or falls within a difference threshold of the second value of the historical data associated with the one or more users.
10 . The computing device of claim 8 , wherein executing the computer-executable instructions further causes the one or more processors to obtain the historical data associated with the one or more users from a second textual data instance of the set of textual data instances based at least in part on identifying that the second textual data instance is associated with a second data type of the set of data types.
11 . The computing device of claim 8 , wherein the historical data associated with the one or more users comprises historical transactions of one or more accounts corresponding to the one or more users.
12 . The computing device of claim 8 , wherein the automated process is one of a plurality of automated processes, and wherein executing the computer-executable instructions further causes the one or more processors to:
identify a second textual data instance of the set of textual data instances based on a second data type identified for the second textual data instance
identify an approved process type based at least in part on a value corresponding to the second textual data instance; and
select the automated process from the plurality of automated processes based at least in part on the approved process type that was identified.
13 . The computing device of claim 8 , wherein the image is associated with an online account, and wherein the online account and the first value corresponding to the textual data instance are utilized to execute the automated process.
14 . The computing device of claim 8 , wherein the image is one of a plurality of images obtained by the computing device over a time period, and wherein the computing device obtains the plurality of images based at least in part on receiving user input indicating a periodicity at which images are to be requested from a second computing device.
15 . A non-transitory computer-readable medium comprising computer-executable instructions that, when executed by one or more processors of a computing device, cause the one or more processors to:
receive an image having an unstructured format;
provide the image as input to a machine-learning model that has been previously trained to identify a set of textual data instances within the image and a set of data types corresponding to the set of textual data instances, the machine-learning model being previously trained utilizing a machine-learning algorithm and a training data set comprising example images of a plurality of unstructured formats, each example identifying corresponding textual data instances within a corresponding image of a respective unstructured format of the plurality of unstructured formats and corresponding data types of the corresponding textual data instances;
obtain, from the machine-learning model, output identifying the set of textual data instances within the image and the set of data types corresponding to the set of textual data instances;
determine whether a textual data instance of the set of textual data instances is valid based at least in part on identifying a data type for the textual data instance and comparing a first value corresponding to the textual data instance to a second value of historical data associated with one or more users; and
based at least in part on determining that the textual data instance is valid, execute operations that cause an automated process to be executed using the one or more values corresponding to the set of textual data instances.
16 . The non-transitory computer-readable medium of claim 15 , wherein executing the computer-executable instructions that determine whether the textual data instance of the set of textual data instances is valid further causes the one or more processors to determine whether the first value corresponding to the textual data instance matches or falls within a difference threshold of the second value of the historical data associated with the one or more users.
17 . The non-transitory computer-readable medium of claim 15 , wherein executing the computer-executable instructions further causes the one or more processors to obtain the historical data associated with the one or more users from a second textual data instance of the set of textual data instances based at least in part on identifying that the second textual data instance is associated with a second data type of the set of data types.
18 . The non-transitory computer-readable medium of claim 15 , wherein the historical data associated with the one or more users comprises historical transactions of one or more accounts corresponding to the one or more users.
19 . The non-transitory computer-readable medium of claim 15 , wherein the automated process is one of a plurality of automated processes, and wherein executing the computer-executable instructions further causes the one or more processors to:
identify a second textual data instance of the set of textual data instances based on a second data type identified for the second textual data instance
identify an approved process type based at least in part on a value corresponding to the second textual data instance; and
select the automated process from the plurality of automated processes based at least in part on the approved process type that was identified.
20 . The non-transitory computer-readable medium of claim 15 , wherein the image is associated with an online account, and wherein the online account and the first value corresponding to the textual data instance are utilized to execute the automated process.