The neural network recognize bad black people and women

Date:

2018-02-15 09:00:04

Views:

170

Rating:

1Like 0Dislike

Share:

The neural network recognize bad black people and women

Facial recognition System already firmly in our lives. It is used in scanners like Face ID and is even used for . But, according to the publication Science Daily, a team of researchers from Massachusetts Institute of technology (MIT) and Stanford, having studied several commercial AI algorithms for face recognition, showed signs of sexism and racism.

The study revealed that in the analysis of photographs of white males, the algorithms incorrectly defined gender only in 0,8% of cases, and in the case of black women, the percentage of errors averaged about 30%. To identify patterns in the employee MIT joy Buolamwini has 1200 photos of different people (which, however, were mostly depicted women with dark skin). Then she turned to the Fitzpatrick scale, according to which there are 6 phototypes of skin, where I is the lightest, and VI being the darkest.

In the end it turned out that women with dark skin with the assessment of IV, V or VI according to Fitzpatrick scale, the error rate was 20.8%, 34.5% and 34.7% respectively. Thus it was possible to establish that the darker the skin phototype, the more the probability that the algorithm will make a mistake. In identifying men with light skin percentage of errors did not exceed 0.8%. Moreover, joy (who is herself African-American) tried to analyze your own photos, and it turned out that the system is either not recognize her face, or incorrectly define gender.

In fact, such research can be treated differently. But putting aside all the jokes and prejudices about tolerance, the study raises a very important question: do the algorithms for facial recognition to work properly? After all, with their help China can already try to prosecute. In addition, the new finding is weakly consistent with the statement of one of the software companies that created software for neural networks and claimed that their software recognize face with accuracy in 97% of cases. What do you think? Express your opinion in our new .

Recommended

Segways have become a cause of the road collapse in California

Segways have become a cause of the road collapse in California

the use of the electric vehicle, without a doubt, a very useful and progressive technology. Especially if we are talking about compact vehicles like scooters designed to relieve the road. However, the use of new vehicles has only positive aspects. Fo...

The neural network recognize bad black men and women

The neural network recognize bad black men and women

facial recognition System already firmly in our lives. It is used in scanners like Face ID and is even used for . But, according to the publication Science Daily, a team of researchers from Massachusetts Institute of technology (MIT) and Stanford, ha...

Comments (0)

This article has no comment, be the first!

Add comment

Related News

RFID tags can be applied to food by using edible graphene

RFID tags can be applied to food by using edible graphene

— an extremely thin and durable structure, consisting of the words of carbon atoms. In addition, graphene is an excellent conductor of heat and electricity and has antibacterial properties. In recent times we often post new...

Artificial intelligence has become to learn 10 times faster and more efficiently

Artificial intelligence has become to learn 10 times faster and more efficiently

a Division of Google, dealing with the development of artificial intelligence, announced the creation of a new method of training neural networks, combining the use of advanced algorithms and old video games. The quality of the le...

Qualcomm continues to prepare for the future with 5G speeds

Qualcomm continues to prepare for the future with 5G speeds

In the coming years, high speed mobile data will become a common occurrence that should be taken to a new level the Internet of things and create new lines of business. As said the head of Qualcomm Steve Mollenkopf, in the near fu...