Russian organisations harness artificial intelligence

Preventing fraudulent transactions

Internally, Yandex itself uses AI and machine learning to prevent fraudulent transactions. The machine learning model analyses transactions and returns a response defining the probability of the transaction being fraudulent.

The company is also using AI to automate bonuses to users in its online payment division, Yandex.Money. A spokesperson for Yandex.Money said: “Now users receive discounts and bonuses based on their interests. To match offers to their interests, a crypto-technology based on machine learning is used, which singles out groups of users due to their behavior online.”

AI on the road

Heavy industry in Russia is also on the look-out for opportunities to use AI. In line with global trends, Russian automotive company KamAZ is working on developing a self-driving truck.

AI solutions for the project are provided by the Moscow-based company Cognitive Technologies. “We have implemented a series of technologies that imitate human vision and cognitive processes,” Olga Uskova, president of Cognitive Technologies group, told Computer Weekly.

She added that the company developed a technology that helps to improve the detection of objects improving road safety thanks to combination of input data from cameras and other sensors at a low level. “Overall, our technologies based on neural networks ensure a detection level of objects of up to 99.9%,” said Uskova.

According to Uskova, Cognitive Technologies’ AI algorithms for KamAZ’s self-driving truck take into account substantial possible vertical fluctuations of the truck, as well as situations of limited vision, such as rain, fog and snowfall, and transitions from light into dark and back. A fully self-driving truck is expected to be developed by 2025.

Applying AI to food processing

Meanwhile, Damate, Russia’s largest producer of turkey meat, is planning to automate and transfer control of up to 80% of its production lines to AI in the next three years.

It is currently evaluating a computer vision system, performing quality control and a system that performs sanitary control by means of video surveillance and computer vision.

The company is also running a pilot project for a system, which checks if the poultry facilities operators are in compliance with their schedules and regulations, known as The Reckoner.

AI solutions for Damate are provided by Connectome.ai, a Russian tech start-up.

“The Reckoner is an automated visual evaluation system that allows quality control on production line,” Connectome.ai CEO Aleksey Hahunov told Computer Weekly. “It requires only a special surveillance camera installed at the quality control gates.”

Another solution developed by Connectome.ai is The Beholder. This is an action recognition system that provides random automatic control over the execution of tasks and using video surveillance checks compliance with the employee’s schedule.

According to Hahunov, the key element of the system is the tracking algorithm that detects what task is performed by an employee and precisely at what location. The algorithm can use the client’s employee database to access profile photos, which allows it to automatically detect who is performing a specific task.

Innovation in banking

In the financial sector, state-run Sberbank has earned a reputation for innovation. In late 2016, it launched a programme for automated decisions on loans which are based on AI.

Earlier this year, another AI-based feature, My Assistant, was made available in the lender’s mobile app, Sberbank Online. It analyses users’ transactions to provide them with the most relevant information related to banking products and general tips, such as tax benefits and discounts on goods and services.

Sberbank said it hopes all 20 million of active users of the mobile app will be using the feature.

Meanwhile, more traditional places for innovation are also responding to growing demand for AI solutions from the economy.

Skolkovo Innovation Centre, Russia’s main IT hub, recently said it will create a specific department for artificial intelligence research.