Image Score: How to Select Useful Samples
This work addresses the challenge of sample selection for improving model performance in semi-supervised learning, though it appears incremental as it builds on existing interpretation methods.
The paper tackles the problem of interpreting neural networks by focusing on why models are error-prone with low softmax scores, proposing an approach to assign scores to samples based on activation chain frequency, and demonstrates that this method can select useful samples to improve deep neural networks in semi-supervised learning.
There has long been debates on how we could interpret neural networks and understand the decisions our models make. Specifically, why deep neural networks tend to be error-prone when dealing with samples that output low softmax scores. We present an efficient approach to measure the confidence of decision-making steps by statistically investigating each unit's contribution to that decision. Instead of focusing on how the models react on datasets, we study the datasets themselves given a pre-trained model. Our approach is capable of assigning a score to each sample within a dataset that measures the frequency of occurrence of that sample's chain of activation. We demonstrate with experiments that our method could select useful samples to improve deep neural networks in a semi-supervised leaning setting.