Understanding Deep Learning Performance through an Examination of Test Set Difficulty: A Psychometric Case Study
This work addresses the challenge of interpreting deep learning performance beyond accuracy for researchers, though it is incremental in applying existing psychometric methods to AI evaluation.
The paper investigates how test set question difficulty, modeled using psychometric methods, affects deep learning model performance, finding that easier examples are learned faster than harder ones as training data increases.
Interpreting the performance of deep learning models beyond test set accuracy is challenging. Characteristics of individual data points are often not considered during evaluation, and each data point is treated equally. We examine the impact of a test set question's difficulty to determine if there is a relationship between difficulty and performance. We model difficulty using well-studied psychometric methods on human response patterns. Experiments on Natural Language Inference (NLI) and Sentiment Analysis (SA) show that the likelihood of answering a question correctly is impacted by the question's difficulty. As DNNs are trained with more data, easy examples are learned more quickly than hard examples.