The Artificial Intelligence (AI) systems have shown that it has been biased towards the people of color. Therefore, the AI systems have been criticized a lot because of this discrimination between white people and black people. For instance, the commercial facial recognition software has been failing and the analysis shows that they are unable to classify people of color. Therefore, Facebook is now sharing a new and diverse dataset of people all over the world to help AI researches to overcome this issue.
This dataset that is provided by Facebook is providing AI with the data of diverse paid actors who were directly asked to share their ages and gender. According to the announcement on Thursday, it is hoped that the research of the AI will use this set of data and will judge whether the AI software is working properly and can distinguish between different ages, gender, races, and skin tones.
The open data that is shared by Facebook to the AI is called Casual Conversations. Facebook describes it as a good and bold first step to help the researchers of AI to identify different people of different skin tones. The Casual Conversation includes about 3,011 people all across the United States and also includes about 45,186 videos. This set of data is names Casual Conversations because the people were recorded while questions were answered casually without being scripted.
Trained professionals were also hired by the company to label the ambient lighting and skin tones of people from different regions according to the Fitzpatrick scale that classify the different skin tones of human. This will help Artificial Intelligence to distinguish between the different people of color.
In this set of data that is provided to AI by Facebook, many of the people participated by their own will, but some of the people are unaware of their participation in the data. Facebook used the ImageNet that is an enormous site for the data of all kinds of images that are gathered from all over the internet. This is the data that is provided to the AI, to make Artificial Intelligence fairer.
The data that is on the Casual Conversation is the same data with the same paid actors that Facebook used for the creation of Deepfake videos. The Facebook company also states that some of the information has been added in the data that was used by the Casual Conversation to make the data of Deepfake data set. Facebook also admits that this is not the perfect data but in the upcoming years it plans to make the data more inclusive and diverse.
Read next: Even after the launch of so many popular social media apps, Facebook and YouTube still remain to be the most widely used platforms
This dataset that is provided by Facebook is providing AI with the data of diverse paid actors who were directly asked to share their ages and gender. According to the announcement on Thursday, it is hoped that the research of the AI will use this set of data and will judge whether the AI software is working properly and can distinguish between different ages, gender, races, and skin tones.
The open data that is shared by Facebook to the AI is called Casual Conversations. Facebook describes it as a good and bold first step to help the researchers of AI to identify different people of different skin tones. The Casual Conversation includes about 3,011 people all across the United States and also includes about 45,186 videos. This set of data is names Casual Conversations because the people were recorded while questions were answered casually without being scripted.
Trained professionals were also hired by the company to label the ambient lighting and skin tones of people from different regions according to the Fitzpatrick scale that classify the different skin tones of human. This will help Artificial Intelligence to distinguish between the different people of color.
In this set of data that is provided to AI by Facebook, many of the people participated by their own will, but some of the people are unaware of their participation in the data. Facebook used the ImageNet that is an enormous site for the data of all kinds of images that are gathered from all over the internet. This is the data that is provided to the AI, to make Artificial Intelligence fairer.
The data that is on the Casual Conversation is the same data with the same paid actors that Facebook used for the creation of Deepfake videos. The Facebook company also states that some of the information has been added in the data that was used by the Casual Conversation to make the data of Deepfake data set. Facebook also admits that this is not the perfect data but in the upcoming years it plans to make the data more inclusive and diverse.
Read next: Even after the launch of so many popular social media apps, Facebook and YouTube still remain to be the most widely used platforms
