Unleashing AI in the Art Market with Roman Kr

image: Roman Kräussl – University of Luxembourg
see Following

Credit: University of Luxembourg

Train a neural network to evaluate art

In this study, which combines the fields of finance and computer science, researchers used machine learning and artificial intelligence to create a neural network algorithm that mimics the work of human evaluators by generating price predictions for art at auction. This algorithm relies on data using both visual and non-visual characteristics of the artwork. The authors of this study unleashed their algorithm on a large art sales dataset capturing 1.2 million painting auctions from 2008 to 2014, training the neural network with both an image of the artwork and information such as artist, medium and auction. house where the work was sold. Once trained on this dataset, the authors asked the neural network to predict the auction house’s pre-sale estimates, the purchase price (the minimum price at which the artwork will sell), as well as the final auction price for the art. sales in 2015. It then became possible to compare the algorithm’s estimate with the actual data and determine whether the relative level of the machine-generated price predictions predicts the relative price results.

The way to a more efficient market?

Not surprisingly, the human experts’ predictions were more accurate than the algorithm, whose prediction, in turn, was more accurate than the standard linear hedonic model the researchers used to compare the study. The reasons for the gap between man and machine include, as the authors argue, primarily access to a greater amount of information about individual works of art, including provenance, condition and the historical context. While interesting, the authors’ goal was not to pit man against machine on this specific task. Rather, the authors aimed to uncover the usefulness and potential applications of machine-based assessments. For example, using such an algorithm, it may be possible to determine whether an auctioneer’s pre-sale assessments are overly pessimistic or overly optimistic, effectively predicting the auctioneers’ prediction errors. Ultimately, this information could be used to correct these types of man-made market inefficiencies.

Beyond the auction block

The implications of this methodology and the computing power applied are not, however, limited to the art world. Other markets trading “real” assets, which rely heavily on human evaluators, namely the real estate market, can benefit from research. Although AI is not yet likely to replace humans, machine learning technology as demonstrated by researchers can become an important tool for investors and intermediaries, who want to access so much information, so quickly. and as cheaply as possible.

Warning: AAAS and EurekAlert! are not responsible for the accuracy of press releases posted on EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.

Back To Top