Sberbank partners with Microsoft in robotics area
The project is set to boost labor productivity when counting coins.
Sberbank Rossii PAO (MCX:SBER) and Microsoft Research (MSR) have announced a joint project that will explore the use of Microsoft AI solutions in robotics. The goal of this project is to train robots to interact with physical objects the way humans do.
Research engineers from Sberbank will be working alongside their MSR counterparts in Redmond, the United States. The team will be looking for the best ways to control manipulators leveraging the capabilities of Microsoft’s AI platform, which unites machine learning – including reinforcement learning – and simulation modeling in virtual reality in order to create autonomous AI systems capable of operating in real world. The project relies on Sberbank’s robotics lab findings regarding precise manipulation of objects that have different and changing shapes by using computer vision and deep learning.
The project is expected to result in an updated control system to operate the robotics unit that is intended to boost labor productivity when counting coins. Featuring a manipulator, video cameras, and an arm grabber, the solution is poised to remove toil and improve occupational safety of the operators who remove coin bags from carts when working at cash handling and cash-in-transit centers, thus improving the bank’s efficiency.
The findings of the joint study are expected to be later used in writing academic papers on the practical use of artificial intelligence.
Sberbank has long been known as a keen adopter of AI solutions. In February this year, for instance, the bank announced the launch of an AI-based chatbot that will help users of electronic trading system Sberbank Markets.
Speaking of robotic solutions that assist in money counting, let’s recall that, in May 2017, Sberbank showed a “Kuka Hand” – a robotic hand used to sort and count money.
