TRAINING THE CNN. The Tokyo team's AI is what's known as a convolutional neural network (CNN), a kind of AI frequently used to analyze images. To train it, the researchers used an automated particle analyzer to generate thousands of two-dimensional images of ash particles.
Then, they manually classified the particles into four base shapes: blocky, vesicular, elongated, or rounded. Some particles were a combination of shapes. Next, they fed their CNN 200 images of particles that fit neatly into each category, using those images to teach it to categorize the particles based on certain parameters (for example, rounded particles should have a high degree of circularity).