☰
✕
UA RU EN

How Artists’ Intuition and the ResMem AI Predict What Images Stick in Memory

How Artists’ Intuition and the ResMem AI Predict What Images Stick in Memory Photo: НВ — Техно

Exploring Artistic Intuition in Chicago

A study conducted at the University of Chicago revealed that artists have an innate ability to produce images that audiences remember more vividly. Over a three-month period, researchers gathered 86 artworks from various artists, selecting 40 pieces for a special exhibition: 20 that were highly memorable and 20 that were less so. The exhibition was held at a Chicago gallery, allowing visitors to experience firsthand the findings of the research.

Key Findings from the Experiment

Participants in the study consciously created images during an online test, aiming either for maximum recognizability or for designs that would be easily forgotten. The results confirmed that artworks intended to be unforgettable were indeed recalled better by viewers. This effect was particularly strong for simpler images that featured fewer repeating elements and less symmetry.

A groundbreaking part of the research involved the ResMem neural network, which had previously been trained exclusively on photographs rather than artistic creations. Remarkably, ResMem outperformed both the participating artists and average viewers in predicting which images would be most memorable. The study’s findings were published in the journal Psychological Science. This highlights the crucial role of an artist’s intuition, as their internal sense guides them in crafting works that resonate and linger in the audience’s mind.

These insights hold significant implications for the future of contemporary art and art education. Understanding how to make images memorable could empower artists to connect more deeply with their audiences and amplify the impact of their work. Additionally, incorporating advanced technologies like neural networks into artistic processes opens new avenues for exploring the intersection of creativity and scientific innovation.