Memorization in deep learning: A survey

TL;DR

This survey reviews the phenomenon of memorization in deep neural networks, exploring its implications for generalization, security, and privacy, and discusses evaluation methods, vulnerabilities, and applications related to memorization.

Contribution

It provides the first comprehensive framework and literature review on DNN memorization, connecting it to generalization, security, and privacy issues, and highlighting applications and challenges.

Findings

Memorization affects model generalization and security.

Evaluation methods exist at example and model levels.

Memorization can be exploited for privacy preservation.

Abstract

Deep Learning (DL) powered by Deep Neural Networks (DNNs) has revolutionized various domains, yet understanding the intricacies of DNN decision-making and learning processes remains a significant challenge. Recent investigations have uncovered an interesting memorization phenomenon in which DNNs tend to memorize specific details from examples rather than learning general patterns, affecting model generalization, security, and privacy. This raises critical questions about the nature of generalization in DNNs and their susceptibility to security breaches. In this survey, we present a systematic framework to organize memorization definitions based on the generalization and security/privacy domains and summarize memorization evaluation methods at both the example and model levels. Through a comprehensive literature review, we explore DNN memorization behaviors and their impacts on security and privacy. We also introduce privacy vulnerabilities caused by memorization and the phenomenon of forgetting and explore its connection with memorization. Furthermore, we spotlight various applications leveraging memorization and forgetting mechanisms, including noisy label learning, privacy preservation, and model enhancement. This survey offers the first-in-kind understanding of memorization in DNNs, providing insights into its challenges and opportunities for enhancing AI development while addressing critical ethical concerns.