A widely accepted claim (e.g., 'gradient descent is commonly used to optimize neural networks')

A claim that readers might question

A factual claim that is easily checked

A claim showing controversy in the field

A method standard in the field (e.g., ReLU, k-means)