Abstract:Online continual learning for image classification studies the problem of learning to classify images from an online stream of data and tasks, where tasks may include new classes (class incremental) or data nonstationarity (domain incremental). One of the key challenges of continual learning is to avoid catastrophic forgetting (CF), i.e., forgetting old tasks in the presence of more recent tasks. Over the past few years, a large range of methods and tricks have been introduced to address the continual learning problem, but many have not been fairly and systematically compared under a variety of realistic and practical settings. To better understand the relative advantages of various approaches and the settings where they work best, this survey aims to (1) compare state-of-the-art methods such as Maximally Interfered Retrieval (MIR), iCARL, and GDumb (a very strong baseline) and determine which works best at different memory and data settings as well as better understand the key source of CF; (2) determine if the best online class incremental methods are also competitive in the domain incremental setting; and (3) evaluate the performance of 7 simple but effective tricks such as the "review" trick and the nearest class mean (NCM) classifier to assess their relative impact. Regarding (1), we observe that iCaRL remains competitive when the memory buffer is small; GDumb outperforms many recently proposed methods in medium-size datasets and MIR performs the best in larger-scale datasets. For (2), we note that GDumb performs quite poorly while MIR — already competitive for (1) — is also strongly competitive in this very different (but important) continual learning setting. Overall, this allows us to conclude that MIR is overall a strong and versatile online continual learning method across a wide variety of settings. Finally for (3), we find that all tricks are beneficial, and when augmented with the "review" trick and NCM classifier, MIR produces performance levels that bring online continual learning much closer to its ultimate goal of matching offline training. Our codes are available at https://github.com/RaptorMai/online-continual-learning.

Preliminary Exploration of Data Incremental Learning Method

Class-Incremental Learning: Survey and Performance Evaluation on Image Classification

An Appraisal of Incremental Learning Methods

Incremental learning with neural networks for computer vision: a survey

Incremental Learning Without Looking Back: a Neural Connection Relocation Approach

Online continual learning in image classification: An empirical survey

Class Incremental Learning with Task-Specific Batch Normalization and Out-of-Distribution Detection

General Incremental Learning with Domain-aware Categorical Representations

Incremental Learning In Online Scenario

ICFD: an Incremental Learning Method Based on Data Feature Distribution.

Incremental Learning from Stream Data

Deep Class Incremental Learning from Decentralized Data

Class-Incremental Learning: A Survey

Incremental Scene Classification Using Dual Knowledge Distillation and Classifier Discrepancy on Natural and Remote Sensing Images

Self-Supervised Learning Aided Class-Incremental Lifelong Learning

Error-Driven Incremental Learning in Deep Convolutional Neural Network for Large-Scale Image Classification

Learning to Classify With Incremental New Class

Large Scale Incremental Learning

Class Incremental Learning Based on Local Structure Constraints in Feature Space

Incremental Learning Using Conditional Adversarial Networks

Class Incremental Learning with Self-Supervised Pre-Training and Prototype Learning