Sberbank's neural networks learn lending while playing poker

According to the head of Sberbank German Gref, the organization intends to apply artificial intelligence for lending to medium and large businesses. For better risks calculating and bluff determining, the bank’s AI algorithm is trained during playing poker with real people.

Neural networks show high results in such poker competitions. The management of Sberbank also notes that AI systems have already proved their effectiveness in granting loans to individuals and small businesses. They have the lowest rate of overdue.

If artificial intelligence can improve the subjective factor through poker, it can be used for making large financial deals in about 3 years.

Subscribe to the news of RGW Moscow in Facebook, VK and Twitter!