Visual representation generation for bias correction
In some implementations, a device may determine an interaction profile for a provider that is to engage with a user in a communication session. The interaction profile may be based on interaction data relating to interpersonal interactions involving the provider during one or more previous communication sessions. The interaction profile may indicate a bias of the provider in connection with one or more categories of users. The device may generate, based on the interaction profile, a visual representation that depicts at least a face of a person for presentation to the provider during the communication session. One or more characteristics associated with the one or more categories of users may be absent from the face of the person. The device may cause presentation of the visual representation to the provider during the communication session.
1 . A system for image generation, the system comprising:
one or more memories; and
one or more processors, communicatively coupled to the one or more memories, configured to:
determine, using a machine learning model and based on interaction data, an interaction profile for a provider,
wherein the interaction data includes demographic information of users involved in interactions with the provider, and
wherein the interaction profile provides an indication of bias of the provider toward one or more categories of users;
generate an image that depicts at least a face of a person,
wherein one or more characteristics associated with the one or more categories of users are absent from the face of the person; and
cause presentation of the image on a device associated with the provider during a communication session with a user.
2 . The system of claim 1 , wherein the interaction data further includes at least one of:
transcripts for one or more previous communication sessions involving the provider, or
reviews of the one or more previous communication sessions.
3 . The system of claim 1 , wherein the interaction profile is unique to the provider.
4 . The system of claim 1 , wherein the one or more processors, to determine the interaction profile, are configured to:
input, into the machine learning model, the interaction data to cause the machine learning model to output the interaction profile.
5 . The system of claim 1 , wherein the machine learning model is a first machine learning model; and
wherein the one or more processors, to generate the image, are configured to:
generate the image using a second machine learning model.
6 . The system of claim 5 , wherein the one or more processors, to generate the image, are configured to:
input, into the second machine learning model, the interaction profile to cause the second machine learning model to output the image.
7 . The system of claim 1 , wherein the communication session comprises at least one of a voice call or an electronic chat.
8 . A non-transitory computer-readable medium storing a set of instructions, the set of instructions comprising:
one or more instructions that, when executed by one or more processors of a device, cause the device to:
determine, using a machine learning model and based on interaction data, an interaction profile for a provider,
wherein the interaction data includes demographic information of users involved in interactions with the provider, and
wherein the interaction profile provides a binary indication of bias of the provider toward one or more categories of users;
generate an image that depicts at least a face of a person,
wherein one or more characteristics associated with the one or more categories of users are absent from the face of the person; and
cause presentation of the image on the device during a communication session with a user.
9 . The non-transitory computer-readable medium of claim 8 , wherein the one or more categories of users include:
an age category,
a male category,
a female category,
a physical trait category,
an occupation category, or
a financial category.
10 . The non-transitory computer-readable medium of claim 8 , wherein the interaction data further includes transcripts for one or more previous communication sessions involving the provider.
11 . The non-transitory computer-readable medium of claim 8 , wherein the interaction profile is unique to the provider.
12 . The non-transitory computer-readable medium of claim 8 , wherein the machine learning model is a first machine learning model; and
wherein the one or more instructions, that cause the device to generate the image, cause the device to:
generate the image using a second machine learning model.
13 . The non-transitory computer-readable medium of claim 12 , wherein the one or more instructions, that cause the device to generate the image, cause the device to:
input, into the second machine learning model, the interaction profile to cause the second machine learning model to output the image.
14 . A method, comprising:
determining, by a device using a machine learning model and based on interaction data, an interaction profile for a provider,
wherein the interaction data includes demographic information of users involved in interactions with the provider, and
wherein the interaction profile provides an indication of bias of the provider toward one or more categories of users;
generating, by the device, a visual representation that depicts at least a face of a person,
wherein one or more characteristics associated with the one or more categories of users are absent from the face of the person; and
causing, by the device, presentation of the visual representation on the device during a communication session with a user.
15 . The method of claim 14 , wherein the interaction profile is unique to the provider.
16 . The method of claim 14 , wherein the indication of bias is an indication of a degree of bias.
17 . The method of claim 14 , wherein determining the interaction profile comprises:
inputting, into the machine learning model, the interaction data to cause the machine learning model to output the interaction profile.
18 . The method of claim 17 , wherein the machine learning model is a first machine learning model; and
wherein generating the visual representation comprises:
inputting, into a second machine learning model, the interaction profile to cause the second machine learning model to output the visual representation.
19 . The method of claim 14 , wherein the visual representation is one of:
an image,
a video, or
a three-dimensional model.
20 . The method of claim 14 , wherein the communication session comprises at least one of a voice call or an electronic chat.