Sound detection alerts
Custom alerts may be generated based on sound type indicators determined using a machine learning classification model trained on user-provided sound recordings and user-defined sound type indicators. A device may provide a sound recording and a type indicator identifying an entity that made a sound in the recording for storage in a database that includes a plurality sound recordings associated with a plurality of type indicators. A machine learning classification model may be trained based on the stored recordings, including the user-defined recordings. The model may be used to classify sounds recorded by other devices and generate alerts identifying the type of sound. Thus, multiple users may contribute data to customize machine learning models that recognize sounds and generate alerts based on user-defined identifiers.
1. A computer system, comprising:
one or more processors; and
one or more machine-readable media coupled to the one or more processors and storing computer program code comprising sets of instructions executable by the one or more processors to:
receive a first sound recording data object and a first sound type indicator from a first device, the first sound type indicator identifying a first type of entity that made a first sound represented by the first sound recording data object;
store the first sound recording data object and the first sound type indicator in a database, the first sound type indicator being associated with the first sound recording data object in the database, the database including a plurality of sound recording data objects, each of the plurality of sound recording data objects being associated with one or more of a plurality of sound type indicators, the plurality of sound recording data objects including a set of sound recording data objects representing sounds made by the first type of entity;
provide to a machine learning classification model based on the first sound recording data object from the first device, the first sound type indicator from the first device, the plurality of sound recording data objects, and the one or more of the plurality of sound type indicators associated with the plurality of sound recording data objects, the machine learning classification model configured to receive an input sound recording data object and provide an output sound type indicator;
train a machine learning classification model based on a sound library, user provided sound recording data, user provided sound type indicators, the first sound recording data object from the first device, the first sound type indicator from the first device, the plurality of sound recording data objects, and the one or more of the plurality of sound type indicators associated with the plurality of sound recording data objects;
receive a second sound recording data object from a second device, the second sound recording data object representing a second sound made by the first type of entity;
determine that the second sound recording data object was made by the first type of entity identified by the first sound type indicator by applying the second sound recording data object to the machine learning classification model;
send the first sound type indicator identifying the first type of entity to the second device causing the second device to generate an alert identifying the first type of entity to a user of the second device; and
provide the user through the second device with customized options regarding delivery of the alert as icons, text, visualizations, tactile, or haptic techniques on or by the second device.
2. The computer system of claim 1 , wherein the alert includes one or more of a visual notification by the second device or a haptic notification by the second device.
3. The computer system of claim 1 , wherein the computer program code further comprises sets of instructions executable by the one or more processors to:
preprocess the first sound recording data object before the storing of the first sound recording data object in the database.
4. The computer system of claim 3 , wherein the preprocessing of the first sound recording data object includes performing noise reduction on the first sound recording data object.
5. The computer system of claim 3 , wherein the computer program code further comprises sets of instructions executable by the one or more processors to:
preprocess the second sound recording data object before the applying the second sound recording data object to the machine learning classification model.
6. The computer system of claim 1 , wherein the computer program code further comprises sets of instructions executable by the one or more processors to:
provide the machine learning classification model to the first device or the second device.
7. The computer system of claim 1 , wherein the machine learning classification model is located on a cloud platform and accessed through a computer server.
8. One or more non-transitory computer-readable media storing computer program code comprising sets of instructions to:
receive a first sound recording data object and a first sound type indicator from a first device, the first sound type indicator identifying a first type of entity that made a first sound represented by the first sound recording data object;
store the first sound recording data object and the first sound type indicator in a database, the first sound type indicator being associated with the first sound recording data object in the database, the database including a plurality of sound recording data objects, each of the plurality of sound recording data objects being associated with one or more of a plurality of sound type indicators, the plurality of sound recording data objects including a set of sound recording data objects representing sounds made by the first type of entity;
provide to a machine learning classification model based on the first sound recording data object from the first device, the first sound type indicator from the first device, the plurality of sound recording data objects, and the one or more of the plurality of sound type indicators associated with the plurality of sound recording data objects, the machine learning classification model configured to receive an input sound recording data object and provide an output sound type indicator;
train a machine learning classification model based on a sound library, user provided sound recording data, user provided sound type indicators, the first sound recording data object from the first device, the first sound type indicator from the first device, the plurality of sound recording data objects, and the one or more of the plurality of sound type indicators associated with the plurality of sound recording data objects;
receive a second sound recording data object from a second device, the second sound recording data object representing a second sound made by the first type of entity;
determine that the second sound recording data object was made by the first type of entity by applying the second sound recording data object to the machine learning classification model;
send the first sound type indicator identifying the first type of entity to the second device causing the second device to generate an alert identifying the first type of entity to a user of the second device; and
provide the user through the second device with customized options regarding delivery of the alert as icons, text, visualizations, tactile, or haptic techniques on or by the second device.
9. The non-transitory computer-readable media of claim 7 , wherein the alert includes one or more of a visual notification by the second device or a haptic notification by the second device.
10. The non-transitory computer-readable media of claim 8 , wherein the computer program code further comprises sets of instructions to:
preprocess the first sound recording data object before the storing of the first sound recording data object in the database.
11. The non-transitory computer-readable media of claim 10 , wherein the preprocessing of the first sound recording data object includes performing noise reduction on the first sound recording data object.
12. The non-transitory computer-readable media of claim 10 , wherein the computer program code further comprises sets of instructions to:
preprocess the second sound recording data object before the applying the second sound recording data object to the machine learning classification model.
13. The non-transitory computer-readable media of claim 8 , wherein the computer program code further comprises sets of instructions to:
provide the machine learning classification model to the first device or the second device.
14. The non-transitory computer readable media of claim 8 , wherein the machine learning classification model is located on a cloud platform and accessed through a computer server.
15. A computer-implemented method, comprising:
receiving a first sound recording data object and a first sound type indicator from a first device, the first sound type indicator identifying a first type of entity that made a first sound represented by the first sound recording data object;
storing the first sound recording data object and the first sound type indicator in a database, the first sound type indicator being associated with the first sound recording data object in the database, the database including a plurality of sound recording data objects, each of the plurality of sound recording data objects being associated with one or more of a plurality of sound type indicators, the plurality of sound recording data objects including a set of sound recording data objects representing sounds made by the first type of entity;
providing to a machine learning classification model based on the first sound recording data object from the first device, the first sound type indicator from the first device, the plurality of sound recording data objects, and the one or more of the plurality of sound type indicators associated with the plurality of sound recording data objects, the machine learning classification model configured to receive an input sound recording data object and provide an output sound type indicator;
training a machine learning classification model based on a sound library, user provided sound recording data, user provided sound type indicators, the first sound recording data object from the first device, the first sound type indicator from the first device, the plurality of sound recording data objects, and the one or more of the plurality of sound type indicators associated with the plurality of sound recording data objects;
receiving a second sound recording data object from a second device, the second sound recording data object representing a second sound made by the first type of entity;
determining that the second sound recording data object was made by the first type of entity by applying the second sound recording data object to the machine learning classification model;
sending the first sound type indicator identifying the first type of entity to the second device causing the second device to generate an alert identifying the first type of entity to a user of the second device; and
providing the user through the second device with customized options regarding delivery of the alert as icons, text, visualizations, tactile, or haptic techniques on or by the second device.
16. The computer-implemented method of claim 15 , wherein the alert includes one or more of a visual notification by the second device or a haptic notification by the second device.
17. The computer-implemented method of claim 15 , further comprising:
preprocessing the first sound recording data object before the storing of the first sound recording data object in the database; and
preprocessing the second sound recording data object before the applying the second sound recording data object to the machine learning classification model.
18. The computer-implemented method of claim 17 , wherein the preprocessing of the first sound recording data object includes performing noise reduction on the first sound recording data object.
19. The computer-implemented method of claim 15 , further comprising:
providing the machine learning classification model to the first device or the second device.
20. The computer-implemented method of claim 15 , wherein the machine learning classification model is located on a cloud platform and accessed through a computer server.