Unsupervised Difficulty Estimation with Action Scores
This work addresses the problem of evaluating difficulty and biases in machine learning models for practitioners deploying models in real-world situations, offering an incremental approach.
This paper introduces a method called 'action score' to estimate the difficulty of individual samples during model training by accumulating losses. The method is unsupervised and requires no model modification, providing insights into model and dataset biases in image classification and object detection.
Evaluating difficulty and biases in machine learning models has become of extreme importance as current models are now being applied in real-world situations. In this paper we present a simple method for calculating a difficulty score based on the accumulation of losses for each sample during training. We call this the action score. Our proposed method does not require any modification of the model neither any external supervision, as it can be implemented as callback that gathers information from the training process. We test and analyze our approach in two different settings: image classification, and object detection, and we show that in both settings the action score can provide insights about model and dataset biases.