Gemtelligence: artificial intelligence is a gem’s best friend
Fake or fortune: uncovering the value of gemstones
The gemstone industry is a treasure trove of sparkling allure and brilliant worth, making gemstones a unique and valuable investment. To understand gemstones’ financial value, experts must first accurately identify and authenticate them. Professional gemstone authentication is a long-established process, and it can take up to six years to train a gemologist in the trade’s intricacies, including determining a gemstone’s type, origin, or possible treatment.
Today, the Gübelin Gem Lab uses Gemtelligence, the artificial intelligence (AI) technology it developed in collaboration with CSEM in its authentication processes to complement its human experts. Gemtelligence uses deep learning models trained on multiple data sets to help Gübelin’s experts determine gemstones’ identity, authenticity, and country of origin with improved consistency.
Gübelin Gem Lab
CSEM’s Gemtelligence deep learning model has enabled Gübelin Gem Lab to significantly improve our results’ consistency and services efficiency, resulting in even greater consumer confidence in our work. By speeding up our workflows, the data-driven AI solution helps us unlock new and valuable revenue streams.
The role of AI in quality control
Gemtelligence was created by feeding huge quantities of data into a CSEM-developed deep learning model. The model provides systematic and reliable insights about colored gemstones, which are used by Gübelin’s experts to support their gemstone evaluations.
The suite was trained with Gübelin’s 28,000-stone reference collection, which contains scientific data on gemstones with provenance and more than 150,000 high-value customer stones. This data has been compiled over the course of 100 years, since the gem laboratory was founded. It includes analysis results from optical and chemical spectrometers, as well as information identifying the characteristics of heated or clarity-enhanced stones. Uniquely, the collection also contains Gübelin experts’ handwritten descriptions and subjective observations.
It’s like having a forensic scientist in the room analyzing and evaluating all the evidence to support your conclusions, which it then learns from to improve from case to case.
As a result, Gemtelligence is helping the Gübelin Gem Lab reduce its experts’ workload by relieving them of some of gemology’s more time-consuming analytical aspects. Crucially, however, the AI solution ensures experts’ gem evaluation quality and consistency, as a key feature of AI technologies in general is that they help eliminate human bias or error.
Exploring unknown possibilities
Tailored in this case to the gemstone industry, the architecture behind Gemtelligence could be applied to an ever-growing list of industries. The deep learning model has the potential to be adapted to any sector where productivity or workflow optimization is needed, which in turn can bring quality-assured products to the market faster than ever before.
As advanced digital technologies evolve, so do the businesses that can benefit from them. Connecting businesses with technological solutions, CSEM’s strength lies in developing AI solutions for every industrial challenge, from the gemstone industry and beyond.