Russian startup FscoreLab has launched a free demo version of the eponymous service for assessing people’s creditworthiness, including neural network analysis of photos.
The service uses a neural network trained on 600 thousand real cases for the payment of loans taken from the database of credit bureaus. The program studied photos of people with and without overdue loans and independently identified the features peculiar to the persons of debtors, follows from the description of the service.
Read the full version on the website rb.ru