Do you like the thick brushstrokes and soft palette of impressionist paintings like Claude Monet? Or do you like Rosco’s bold colors and abstract shapes? Although individual artistic tastes hold secrets, a new study by the California Institute of Technology shows that a simple computer program can predict exactly which painting a person will like.
New research results published in the journal Natural human behaviorUsing Amazon’s Mechanical Turk crowdsourcing platform with more than 1,500 volunteers to rate paintings in the Impressionism, Cubism, Abstract and Colorfield genres.
The volunteers ‘responses were entered into a computer program, and after this familiarization period the computer was able to predict the volunteers’ artistic preferences much better than it happened by chance.
“I was surprised by this result because I thought the evaluation of art was personal and subjective,” said Kiyohito Iigaya, a postdoctoral fellow who works in the lab of John Odhati, a professor of psychology at the California Institute of Technology. Says.
The results not only showed that computers can make these predictions, but also led to a new understanding of how people judge art.
“It is important that we have insight into the mechanisms by which people make aesthetic decisions,” says Odherty. “That is, people seem to combine them with basic pictorial features. This is the first step in understanding how the process works. “
In this study, the team will break down the visual attributes of a painting into low-level features (properties like contrast, saturation, hue, etc.) and high-level features that require human judgment and include: I programmed the computer. Features such as whether the painting is dynamic or stationary.
Computer programs then estimate how much a particular feature is taken into account when deciding how much you like a particular work of art. These decisions combine both low-level and high-level features. If the computer appreciates it, it can successfully predict the preference for another work of art that has not been seen before. “
Kiyohito Iigaya, first author, postdoctoral fellow
Researchers have also found that volunteers tend to fall into three general categories. People who like paintings with real objects, such as impressionist paintings. Someone who likes colorful abstract paintings like Rosco. People who like complex paintings, like Picasso’s Cubist portraits. The vast majority of people fell into the first “real” category. “A lot of people liked Impressionist paintings,” says Iigaya.
In addition, researchers have found that Deep Convolutional Neural Networks (DCNNs) can be trained to learn to predict the artistic preferences of volunteers with similar accuracy. DCNN is a type of machine learning program that provides a computer with a series of training images so you can learn to classify objects such as cats and dogs. These neural networks have interconnected units, like neurons in the brain. Networks can “learn” by changing the strength of the connection from one device to another.
In this case, the deep learning approach did not include the selected low or high level visual features used in the first part of the study, so the computer had to “determine” the features to be analyzed. would have.
“Deep neural network models learn just like the real brain, so we cannot know exactly how the network solves a particular task,” explains Iigaya. Do. “It may be very strange, but when I looked into the neural network, I found that it built the same functional category as the one I chose.” These results determine aesthetic preferences. It suggests that the functions used for this can occur naturally in a brain-like architecture.
“We are now actively investigating whether this is the case by examining people’s brains while people are making the same decisions,” says Odherty.
In another part of the study, the researchers also showed that their simple computer program, already trained in artistic tastes, could predict exactly which photo volunteers would like.
They showed volunteers pictures of scenes like pools and food and saw results similar to paintings. In addition, researchers have shown that reversing the order is also effective. After initially training the volunteers in photography, they were able to use the program to accurately predict the artistic preferences of the subject.
Computer programs have managed to predict the artistic tastes of volunteers, but researchers say there is still more to learn about the nuances that go into personal tastes.
“There are aspects of preference that are unique to a particular person and that could not be successfully explained using this method,” says O’Doherty.
“This more peculiar element may be related to semantic characteristics or other individual personal characteristics that can influence the meaning, past experience and evaluation of the painting. Computer model. It may be possible to identify and learn these traits, but this is a personal preference that may not be generalized to individuals as we have noted here. Includes a more detailed study on. “
California Institute of Technology
Kenichi Iigaya, et al.. (2021) Aesthetic preference for art can be predicted from a mixture of low and high visual characteristics. Natural human behavior.. doi.org/10.1038/s41562-021-01124-6 ..