Task-Free Continual Learning

By Rahaf Aljundi, Klaas Kelchtermans, Tinne Tuytelaars. 2018

432 citations recorded by the source corpus.

Methods proposed in the literature towards continual deep learning typically operate in a task-based sequential learning setup. A sequence of tasks is learned, one at a time, with all data of current task available but not of previous or future tasks. Task boundaries and identities are known at all times. This setup, however, is rarely encountered in practical applications. Therefore we investigate how to transform continual learning to an online setup. We develop a system that keeps on learning over time in a streaming fashion, with data distributions gradually changing and without the notion of separate tasks. To this end, we build on the work on Memory Aware Synapses, and show how this method can be made online by providing a protocol to decide i) when to update the importance weights, ii) which data to use to update them, and iii) how to accumulate the importance weights at each update step. Experimental results show the validity of the approach in the context of two applications: (self-)supervised learning of a face recognition model by watching soap series and learning a robot to avoid collisions.

Subfields: Artificial Intelligence, Computational Mechanics, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computational Mechanics

In Computer Vision and Pattern Recognition

In Representation & Self-Supervised Learning

Citations in the corpus

The Curse of Recursion: Training on Generated Data Makes Models Forget (2023, methodology): “Our work is more similar to task-free continual learning [Aljundi et al., 2019] where data distributions gradually change without the notion of separate tasks.”
A Comprehensive Survey of Continual Learning: Theory, Method and Application (2023, neutral): “Potential challenges of the learning paradigm can be summarized in a broad concept called General Continual Learning (GCL) [12], [20], [31], where the model observes incremental data in an online fashion without explicit task boundaries.”
Task-Free Continual Learning via Online Discrepancy Distance Learning (2022, neutral): “In this paper, we address a more realistic scenario, called Task-Free Continual Learning (TFCL) [3], where task identities are not available while the model can only access a small batch of samples at a given time.”
Continual Learning by Modeling Intra-Class Variation (2022, methodology): “We perform the experiments on CIFAR-100 (5 tasks with 20 classes per task) and TinyImageNet (10 tasks with 20 classes per task) with buffer size 500.”
Few-Shot Continual Active Learning by a Robot (2022, background): “Overall, FoCAL is a combination of multiple challenging problems in machine learning, mainly Few-Shot Class Incremental Learning (FSCIL) [9, 10], Active Learning [14, 18], and online continual learning [19].”
Meta-Learning with Less Forgetting on Large-Scale Non-Stationary Task Distributions (2022, neutral): “Many works mitigate catastrophic forgetting during the learning process [37,14,48,68,31,42,20,2,10,3,64].”
Improving Task-free Continual Learning by Distributionally Robust Memory Evolution (2022, methodology): “Maximally Interfering Retrieval (MIR) (Aljundi et al., 2019a), the goal of MIR is to select the examples that are easily forgettable for replay.”
Simpler is Better: off-the-shelf Continual Learning Through Pretrained Backbones (2022, methodology): “Several methods in CL use pretrained backbones as feature extractors such as in Hayes et al [7] or [1,9] and sometimes the pretraining is carried on half (or a big portion) the dataset considered, as in PODNet [4] or in Yu et al.”
Information-theoretic Online Memory Selection for Continual Learning (2022, neutral): “We also include the feature-space clustering method (FSS-Clust) (Aljundi et al., 2019b) for Permuted MNIST and Split MNIST.”
Continual Normalization: Rethinking Batch Normalization for Online Continual Learning (2022, methodology): “…et al., 2018; von Oswald et al., 2020; Yoon et al., 2018; Li et al., 2019; Xu & Zhu, 2018), experience replay (Lin, 1992; Chaudhry et al., 2019b; Aljundi et al., 2019a; Riemer et al., 2019; Rolnick et al., 2019; Wu et al., 2019), regularization (Kirkpatrick et al., 2017; Zenke et al., 2017;…”
CoMPS: Continual Meta Policy Search (2021, background): “Both terms describe the process of learning tasks in sequence while avoiding the problem of forgetting (French, 1999; Rusu et al., 2015; Parisotto et al., 2016; Rusu et al., 2016; Chen et al., 2015; Tessler et al., 2016; Aljundi et al., 2019).”
Learning curves for continual learning in neural networks: Self-knowledge transfer and forgetting (2021, background): “It would also be interesting to investigate richer but complicated situations required in practice, such as streaming of non-i.i.d. data and distribution shift (Aljundi et al., 2019).”
Online Continual Learning on Class Incremental Blurry Task Configuration with Anytime Inference (2021, methodology): “To address this issue, a task-free setup (Aljundi et al., 2019b), where task-ID at training is not available, has been proposed.”
Recent Advances of Continual Learning in Computer Vision: An Overview (2021, neutral): “Besides VCL [23] and its extensions, various other image classification methods [188,195] have also been proposed.”
Online Continual Learning with Natural Distribution Shifts: An Empirical Study with Visual Data (2021, background): “[2] use soap opera series for actor identification.”
Continual semi-supervised learning through contrastive interpolation consistency (2021, background): “: a robot that explores the environment and learns to avoid collisions by receiving direct feedback from it [3]).”
How Well Does Self-Supervised Pre-Training Perform with Streaming Data? (2021, methodology): “Following the prevailing use of MAS regularization (Aljundi et al., 2018; 2019), we update the MAS importance weights Ω ij to cover information of each data chunk in the sequential training process.”
New Insights on Reducing Abrupt Representation Change in Online Continual Learning (2021, background): “We also explore in this paper a more general setting without the notion of clearly delineated tasks (Aljundi et al., 2018; Chen et al., 2020a), where the data distribution gradually changes over time.”
Rectification-based Knowledge Retention for Continual Learning (2021, background): “The model has to retain the knowledge gained from the older task while learning the new tasks [21, 1, 37].”