Bell Integrator Neuton AutoML

Main articles:

• SaaS - History. Philosophy. Development Drivers
• Machine Learning
• Data management
• Data mining Intelligent data analysis

2021: Entering the Russian market

Bell Integrator On August 3, 2021, the company announced the launch Russian SaaS of the product cloudy , the Neuton AutoML platform, into the market. It allows all categories of users to build predictive models for many business tasks and significantly exceeds other existing solutions in terms of their characteristics, price, deployment time and ease of use.

Neuton AutoML is a universal automated machine learning solution on Google's full-cycle cloud platform (GCP) - from data preparation and training to model deployment and interpretation of prediction results - in a single interface. It was created on the basis of a patented neural network framework (Bell Integrator's own development), which is based on other principles of building neural networks, which allows you to automatically create models with special characteristics.

Unlike other AI solutions, Bell Integrator is so easy to use that it does not require expertise in Data Science, programming, or other specialized expertise. Even business users can independently make forecasts and find insights in the data, whether it be predicting customer outflows, sales forecasts, optimizing business processes, determining the best offer for each client or credit scoring.

We are seeing a boom in the development of artificial intelligence technologies, but all this is done by the hands of machine learning specialists, experts who must combine versatile skills: knowledge of algorithms and mathematics, programming, data science and much more. They are in incredible demand, and many simply cannot afford them. We launched Neuton AutoML to truly democratize artificial intelligence, making it available to all categories of users, "said Ekaterina Vasilenko, product director of Neuton at Bell Integrator.

According to benchmark tests, Neuton AutoML is also different in that it exceeds other developments in such indicators as size (1000 times more compact), calculation speed (100 times faster), implementation speed (it takes days, not months, to implement machine learning using Neuton AutoML) and cost of use. High calculation speed and small size allows you to use models created using Neuton AutoML in microcontrollers, that is, in devices that do not have great speed, and without going to the Internet.

The cloud platform is available on a monthly subscription basis. There is also a free tariff with unlimited opportunities for model training. This option allows users to switch to paid packages, before making sure that the models created solve their business problems. Also, depending on the tariff plan, it is possible to visualize the results of prediction through the web interface, download the ready-to-use model or connect to it through the API for easy integration into the existing business process.

On August 4, 2021, Neuton AutoML launches at Product Hunt, a global platform for the global community of innovators. About 100 products are published on it every day, and they all fight for the attention of the audience.