Enrollment-free offline device personalization

1. A method of personalizing a device, comprising:

receiving first sensor data from one or more sensors;

detecting first biometric information of a user in the first sensor data;

encoding the first biometric information as a first vector using one or more neural network models stored on the device;

matching the first vector to a first entry of a plurality of entries stored in a lookup table, wherein the lookup table is configured to store profile information associated with each entry of the plurality of entries;

incrementing a count value indicating a number of instances the first vector has been matched to the first entry in the lookup table;

determining whether the count value exceeds a threshold number; and

configuring a user interface of the device based at least in part on the profile information associated with the first entry responsive to determining that the count value exceeds the threshold number.

2. The method of claim 1 , wherein the matching comprises:

calculating respective distances from the first vector to each entry of the plurality of entries, including the first entry, stored in the lookup table; and

determining that the distance to the first entry is the shortest among the calculated distances.

3. The method of claim 1 , wherein the profile information for each entry of the plurality of entries includes configurations, settings, preferences, or content to be displayed or rendered via the user interface.

4. The method of claim 1 , further comprising:

monitoring user interactions with the device; and

generating the profile information for each entry of the plurality of entries based at least in part on the user interactions.

5. The method of claim 1 , further comprising:

storing the first vector as a new entry in the lookup table when no matching entry is detected.

6. The method of claim 1 , wherein the first sensor data comprises an image of a scene and the first biometric information comprises one or more facial features of the user in the scene.

7. The method of claim 1 , further comprising:

receiving second sensor data from the one or more sensors;

detecting second biometric information in the second sensor data;

encoding the second biometric information of the second sensor data as a second vector using the one or more neural network models, wherein the second vector is at least a threshold distance away from the first vector; and

changing the user interface based at least in part on the second vector.

8. A device, comprising:

processing circuitry; and

memory storing instructions that, when executed by the processing circuitry, cause the device to:

receive first sensor data from one or more sensors;

detect first biometric information of a user in the first sensor data;

encode the first biometric information as a first vector using one or more neural network models stored on the device;

match the first vector to a first entry of a plurality of entries stored in a lookup table, wherein the lookup table is configured to store profile information associated with each entry of the plurality of entries;

increment a count value indicating a number of instances the first vector has been matched to the first entry in the lookup table;

determine whether the count value exceeds a threshold number; and

configure a user interface of the device based at least in part on the

profile information associated with the first entry responsive to determining that the count value exceeds the threshold number.

9. The device of claim 8 , wherein execution of the instructions for matching the first vector to the first entry causes the device to:

calculate respective distances from the first vector to each entry of the plurality of entries, including the first entry, stored in the lookup table; and

determining that the distance to the first entry is the shortest among the calculated distances.

10. The device of claim 8 , wherein execution of the instructions further causes the device to:

monitor user interactions with the device; and

generate the profile information for each entry of the plurality of entries based at least in part on the user interactions, wherein the profile information for each entry of the plurality of entries includes configurations, settings, preferences, or content to be displayed or rendered via the user interface.

11. The device of claim 8 , wherein execution of the instructions further causes the device to:

store the first vector as a new entry in the lookup table when no matching entry is detected.

12. The device of claim 8 , wherein the first sensor data comprises an image of a scene and the first biometric information comprises one or more facial features of the user in the scene.

13. The device of claim 8 , wherein execution of the instructions further causes the device to:

receive second sensor data from the one or more sensors;

detect second biometric information in the second sensor data;

encode the second biometric information of the second sensor data as a second vector using the one or more neural network models, wherein the second vector is at least a threshold distance away from the first vector; and

change the user interface based at least in part on the second vector.

14. A system comprising:

a memory storing one or more neural network models;

one or more sensors configured to capture at least first sensor data;

a vector generator configured to:

detect first biometric information of a user in the captured first sensor data; and

encode the first biometric information as a first vector using the one or more neural network models stored in the memory;

a profile filter configured to:

match the first vector to a first entry of a plurality of entries stored in a lookup table, wherein the lookup table is configured to store profile information associated with each entry of the plurality of entries;

increment a count value indicating a number of instances the first vector has been matched to the first entry in the lookup table; and

determine whether the count value exceeds a threshold number; and

a display configured to display a user interface based at least in part on the profile information associated with the first entry responsive to determining that the count value exceeds the threshold number.