Enrollment with an automated assistant
Techniques are described herein for dialog-based enrollment of individual users for single- and/or multi-modal recognition by an automated assistant, as well as determining how to respond to a particular user's request based on the particular user being enrolled and/or recognized. Rather than requiring operation of a graphical user interface for individual enrollment, dialog-based enrollment enables users to enroll themselves (or others) by way of a human-to-computer dialog with the automated assistant.
1. A method implemented using one or more processors, comprising:
capturing, at a first local client device, a plurality of images of a person using a vision sensor;
transmitting, from the first local client device and to a remote computing device over one or more networks, the plurality of images;
receiving, at the first local client device and from the remote computing device over one or more of the networks, a message, wherein the message identifies at least one image of the plurality of images as usable to create biometric data that is suitable subsequently to authenticate the person to a computing resource prior to enrolling the person with the computing resource based on the biometric data; and
in response to receiving the message that identifies the at least one image as usable to create the biometric data that is suitable subsequently to authenticate the person to the computing resource:
transmitting, from the first local client device and to a second local client device over one or more of the networks, and based on the message, the at least one image of the plurality of images, wherein transmitting the at least one image of the plurality of images to the second local client device causes the second local client device to generate the biometric data using the at least one image of the plurality of images; and
causing the person to be enrolled with the computing resource based on the biometric data.
2. The method of claim 1 , wherein the biometric data comprises a latent space embedding.
3. The method of claim 2 , wherein the latent space embedding is generated based on application of the at least one image of the plurality of images as input across a machine learning model.
4. The method of claim 1 , wherein the first local client device is a hand held client device of the person.
5. The method of claim 4 , wherein the second local client device is a stationary client device of the person that is co-located in a coordinated ecosystem of client devices along with the hand held client device of the person.
6. The method of claim 5 , wherein causing the person to be enrolled with the computing resource based on the biometric data comprises creating an association between an identity of the person and the biometric data in memory of the stationary client device.
7. The method of claim 1 , wherein the computing resource comprises an automated assistant that semantically processes natural language input to determine an intent, and performs one or more responsive actions based on the intent.
8. A method implemented using one or more processors, comprising:
receiving, from a first remote client device and over one or more networks, a plurality of images of a person;
analyzing the plurality of images to identify at least one image of the plurality of images as usable to create biometric data that is suitable subsequently to authenticate a person to a computing resource prior to the person enrolling with the computing resource based on the biometric data; and
in response to identifying the at least one image as usable to create the biometric data that is suitable subsequently to authenticate the person to the computing resource:
transmitting, to the first remote client device and over one or more of the networks, a message that identifies the at least one image as usable to create biometric data;
wherein the transmitting causes the first remote client device to transmit the at least one image to a second remote client device to generate the biometric data using the at least one image, and
wherein the first remote client device and the second remote client device are co-located in a coordinated ecosystem of client devices.
9. The method of claim 8 , wherein the biometric data comprises a latent space embedding.
10. The method of claim 8 , wherein the analyzing comprises determining that the at least one image of the plurality of images does not capture any faces other than that of the person.
11. The method of claim 8 , wherein the analyzing comprises determining that the at least one image of the plurality of images satisfies one or more quality criterion.
12. The method of claim 8 , wherein the analyzing comprises applying the at least one image of the plurality of images as input across a machine learning model to generate output, wherein the output indicatives that the at least one image is usable to create the biometric data.
13. The method of claim 8 , wherein the computing resource comprises an automated assistant that semantically processes natural language input to determine an intent, and performs one or more responsive actions based on the intent.
14. At least one non-transitory computer-readable medium comprising instructions that, in response to execution by one or more processors, cause the one or more processors to perform the following operations:
capturing, at a first local client device, a plurality of images of a person using a vision sensor;
transmitting, from the first local client device and to a remote computing device over one or more networks, the plurality of images;
receiving, at the first local client device and from the remote computing device over one or more of the networks, a message, wherein the message identifies at least one image of the plurality of images as usable to create biometric data that is suitable subsequently to authenticate the person to a computing resource prior to enrolling the person with the computing resource based on the biometric data; and
in response to receiving the message that identifies the at least one image as usable to create the biometric data that is suitable subsequently to authenticate the person to the computing resource:
generating, at the first local client device, and based on the message, the biometric data using the at least one image of the plurality of image;
causing the person to be enrolled with the computing resource based on the biometric data; and
transmitting, from the first local client device and to a second local client device over one or more of the networks, the biometric data for subsequently authenticating the person to the computing resource.
15. The at least one non-transitory computer-readable medium of claim 14 , wherein the biometric data comprises a latent space embedding.
16. The at least one non-transitory computer-readable medium of claim 15 , wherein the latent space embedding is generated based on application of the at least one image of the plurality of images as input across a machine learning model.
17. The at least one non-transitory computer-readable medium of claim 14 , wherein the first local client device is a hand held client device of the person.
18. The at least one non-transitory computer-readable medium of claim 17 , wherein the second local client device is a stationary client device of the person that is co-located in a coordinated ecosystem of client devices along with the hand held client device of the person.
19. The at least one non-transitory computer-readable medium of claim 18 , wherein causing the person to be enrolled with the computing resource based on the biometric data comprises creating an association between an identity of the person and the biometric data in memory of the stationary client device.
20. The at least one non-transitory computer-readable medium of claim 14 , wherein the computing resource comprises an automated assistant that semantically processes natural language input to determine an intent, and performs one or more responsive actions based on the intent.