13:59, 16 January 2026

Russian Neural Network “Taster” Uses Chemical Fingerprints to Identify Wine Origins

Scientists have developed a digital service that can determine a wine’s region of origin with near-perfect accuracy.

The technology is built around a neural network model capable of processing chemical analysis data and comparing it with a reference database of wine-growing regions. The software analyzes dozens of trace elements identified through laboratory testing and creates a detailed digital profile of each wine. Machine-learning algorithms then match that profile against established “signatures” of known wine-producing areas.

Pattern Recognition

In practice, the system does more than store data. It learns to identify stable patterns, taking into account combinations of elements and their ratios rather than isolated numerical values. This approach makes it possible to distinguish between wines from neighboring regions with nearly identical grape-growing conditions.

“The main outcome of the project is the development of methodological foundations for assessing the origin of wines with protected geographical indications by studying the relationship between multi-element fingerprints and sensory properties,” said Zaualya Temerdasheva, head of the project and chair of analytical chemistry at Kuban State University. “Based on analyses of samples of soils, grapes, and wines from production sites, we identified marker elements that indicate regional and varietal origin for dry white and red wines from Russia’s Krasnodar Territory.”

A “High-Precision” Taster

The system relies on a neural network trained on a large dataset of samples from Russian wine-growing regions. According to Temerdasheva, the accuracy of determining a wine’s origin exceeds 95%.

“Using an application developed with neural network modeling technologies, the elemental composition of a wine is used to establish its match with the elemental signature of a specific region,” she said.

The neural network continues to improve as new data is added. The more wine samples pass through the application, the more precise its conclusions become.

A Digital Wine Lover Takes on Counterfeits

The new technology is designed for practical use. The software can be deployed in quality-control laboratories and used by regulators as well as winemakers themselves. Digital sample comparison simplifies the process of verifying a wine’s origin while delivering highly reliable results.

For producers, the system offers a robust way to protect wines with geographical indications and reduce the risk of counterfeits. Consumers, in turn, gain added confidence that a bottle truly comes from the region stated on the label.

The developers note that the program is not limited to winemaking. The same method could be applied to other products with protected origins, including oils, mineral water, and agricultural raw materials. As a result, the technology could expand beyond a single industry and serve as the foundation for an entire range of digital quality-control services.

Digital products and platforms