Data Cleansing for Models Trained with SGD

TL;DR

This paper introduces a novel data cleansing algorithm for models trained with SGD that identifies influential instances without domain knowledge, improving model accuracy by removing such data points.

Contribution

The paper presents a new method to infer influential data points in SGD-trained models without requiring convex loss functions or optimal models, facilitating easier data cleansing.

Findings

Accurately infers influential instances in SGD-trained models

Improves model accuracy by removing influential data points

Effective on datasets like MNIST and CIFAR10

Abstract

Data cleansing is a typical approach used to improve the accuracy of machine learning models, which, however, requires extensive domain knowledge to identify the influential instances that affect the models. In this paper, we propose an algorithm that can suggest influential instances without using any domain knowledge. With the proposed method, users only need to inspect the instances suggested by the algorithm, implying that users do not need extensive knowledge for this procedure, which enables even non-experts to conduct data cleansing and improve the model. The existing methods require the loss function to be convex and an optimal model to be obtained, which is not always the case in modern machine learning. To overcome these limitations, we propose a novel approach specifically designed for the models trained with stochastic gradient descent (SGD). The proposed method infers the influential instances by retracing the steps of the SGD while incorporating intermediate models computed in each step. Through experiments, we demonstrate that the proposed method can accurately infer the influential instances. Moreover, we used MNIST and CIFAR10 to show that the models can be effectively improved by removing the influential instances suggested by the proposed method.