Signal Strength and Noise Drive Feature Preference in CNN Image Classifiers
This provides guidance for predictive models and bias avoidance in machine vision, but is incremental as it builds on existing feature preference research.
The study investigated feature preference in CNN image classifiers by testing attributes like shape, texture, and color with controlled signal and noise, finding that CNNs prefer features with stronger signal and lower noise regardless of type.
Feature preference in Convolutional Neural Network (CNN) image classifiers is integral to their decision making process, and while the topic has been well studied, it is still not understood at a fundamental level. We test a range of task relevant feature attributes (including shape, texture, and color) with varying degrees of signal and noise in highly controlled CNN image classification experiments using synthetic datasets to determine feature preferences. We find that CNNs will prefer features with stronger signal strength and lower noise irrespective of whether the feature is texture, shape, or color. This provides guidance for a predictive model for task relevant feature preferences, demonstrates pathways for bias in machine models that can be avoided with careful controls on experimental setup, and suggests that comparisons between how humans and machines prefer task relevant features in vision classification tasks should be revisited. Code to reproduce experiments in this paper can be found at \url{https://github.com/mwolff31/signal_preference}.