| The name of the base system (platform): | Artificial intelligence (AI, Artificial intelligence, AI) |
| Developers: | Institute of Artificial Intelligence (III) MIPT, Moscow State University (MSU) |
| Date of the premiere of the system: | 2025/12/15 |
| Technology: | Data Mining |
The main articles are:
2025: Şagaart Development
Scientists at MIPT and Moscow State University have developed an AI model of Şagaart. With machine learning, she calculates the fair value of a work of modern art in seconds. This will help solve the problem of opacity and subjectivity of pricing in the art market. MIPT announced this on December 15, 2025.
How to understand how much a work of art costs? As of December 2025, there is no single set of rules on the art market that would allow a fair price to be calculated. Pricing is chaotic and opaque, creating an atmosphere of distrust, discouraging new buyers and limiting the development of the entire market, the volume of which has not changed on average over the past 15 years and amounts to $60 billion per year.
According to research, only 5% of collectors consider the art market to be completely transparent, and 91% of potential collectors would like to have more information about price formation.
The Şagaart system, developed by scientists at III MIPT and Moscow State University, is designed to solve this problem. She conducts a comprehensive analysis of the work in two stages: first, using algorithmic computer vision, its style, artistic direction and genre are determined. Then the scoring model of machine learning (based on the CatBoost algorithm) conducts an in-depth analysis of market data, taking into account dozens of different parameters: size, material, technique, data on the artist, his exhibition history, the presence of works in museums and private collections, media, detailed sales history, price dynamics, etc. As a result, based on a large array of data, the system calculates the price: reasoned and transparent.
 | The platform will help novice and experienced collectors to test the validity of the price, investors - for decision-making, artists and galleries - to form a reasoned pricing policy, shared Tatyana Shaga, Researcher, Department of Technological Entrepreneurship, MIPT + Skolkovo.
|  |
The technological basis of the system is the modern technology stack: machine models are built on the CatBoost and PyTorch libraries, and the web service core runs on Python and Django, which ensures high reliability and accuracy of calculations.
The important quality of the system is that it saves time. If it usually takes days to analyze a product, Şagaart generates a cost in a few seconds, while providing detailed analytics to support the calculations.
As early as December 2025, the model shows high accuracy: the prediction error after the normalization of the price spread is only 9%. But the effectiveness of the model is still reduced when evaluating lots worth millions of dollars, where uniqueness and subjective perception play a decisive role.
Scientists are actively working to improve algorithms: expand the amount of data and increase accuracy. In the future, developers plan to integrate additional external factors affecting investment demand: key central bank rates, inflation, stock market indices, etc.
| The site content is translated by machine translation software powered by PROMT. The machine-translated articles are not always perfect and may contain errors in vocabulary, syntax or grammar. Read original article If you find inaccuracies or errors in the results of machine translation, please write to editor@tadviser.ru. We will make every effort to correct them as soon as possible. |