Simple Techniques Work Surprisingly Well for Neural Network Test Prioritization and Active Learning (Replicability Study)

TL;DR

This study verifies that simple test prioritization techniques like DeepGini perform comparably to more complex methods in neural network testing, emphasizing the effectiveness of straightforward approaches.

Contribution

The paper provides a large-scale empirical validation showing simple uncertainty-based methods are as effective as complex techniques for neural network test prioritization.

Findings

Simple techniques like DeepGini perform as well as complex methods.

Uncertainty quantification baselines like softmax likelihood are effective.

Complex methods do not significantly outperform simpler baselines.

Abstract

Test Input Prioritizers (TIP) for Deep Neural Networks (DNN) are an important technique to handle the typically very large test datasets efficiently, saving computation and labeling costs. This is particularly true for large-scale, deployed systems, where inputs observed in production are recorded to serve as potential test or training data for the next versions of the system. Feng et. al. propose DeepGini, a very fast and simple TIP, and show that it outperforms more elaborate techniques such as neuron- and surprise coverage. In a large-scale study (4 case studies, 8 test datasets, 32'200 trained models) we verify their findings. However, we also find that other comparable or even simpler baselines from the field of uncertainty quantification, such as the predicted softmax likelihood or the entropy of the predicted softmax likelihoods perform equally well as DeepGini.