Class-Incremental Learning via Dual Augmentation

By Fei Zhu, Zhen Cheng, Xu-Yao Zhang, Chenglin Liu. 2021

244 citations recorded by the source corpus.

Subfields: Meta & Continual Learning

How researchers see this paper

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

In Meta & Continual Learning

1. Xingxing Zhang
2. Han-Jia Ye
3. Liyuan Wang
4. De-chuan Zhan
5. Hang Su
6. Da-Wei Zhou
7. Jun Zhu
8. Bing Liu

Citations in the corpus

• A Comprehensive Survey of Continual Learning: Theory, Method and Application (2023, methodology): “Other methods exploit the class-wise statistics of feature representations to obtain a balanced classifier, such as by imposing the representations to be transferable and invariant (SPB [453], IL2A [509], SSRE [512], FeTrIL [333]), or compensating explicitly the representation shifts (SDC [484], RER”
• S3C: Self-Supervised Stochastic Classifiers for Few-Shot Class-Incremental Learning (2023, background): “Some of the recent works [47, 41, 46] have focused on non-exemplar based methods, with no access to exemplars from the old classes.”
• FeTrIL: Feature Translation for Exemplar-Free Class-Incremental Learning (2022, methodology): “In contrast, IL2A [47], the work which is closest to ours, needs to store a covariance matrix per class to obtain optimal performance.”
• Few-Shot Class-Incremental Learning from an Open-Set Perspective (2022, positive): “Inspired by Mixup [24] and IL2A [26], we randomly combine pairs of different class examples from the base session data to synthesize auxiliary new class data.”
• Deep Class-Incremental Learning: A Survey (2023, methodology): “To this end, exemplar-free CIL [57], [58], [138], [139] should be conducted to enable the model to be adapted without the help of exemplars.”
• Online Continual Learning through Mutual Information Maximization (2022, methodology): “, 2019), ClassAug (Zhu et al., 2021) and our OCM on CIFAR100.”
• Elastic Feature Consolidation for Cold Start Exemplar-free Incremental Learning (2024, methodology): “Most functional regularization approaches use feature distillation (Zhu et al., 2021a; Yu et al., 2020), or even freeze the feature extractor after the first task (Petit et al., 2023; Belouadah & Popescu, 2018; Panos et al., 2023), greatly reducing the plasticity of the network.”
• Divide and not forget: Ensemble of selectively trained experts in Continual Learning (2024, methodology): “Then, the hardest scenario exemplar-free CIL is considered, where number of methods exists: LwF (Li & Hoiem, 2016), SDC (Yu et al., 2020), ABD (Smith et al., 2021), PASS (Zhu et al., 2021b), IL2A (Zhu et al., 2021a), SSRE (Zhu et al., 2022), FeTrIL (Petit et al., 2023).”
• Orchestrate Latent Expertise: Advancing Online Continual Learning with Multi-Level Supervision and Reverse Self-Distillation (2024, methodology): “We compare our approach with 11 OCL methods (AGEM [9], ER [10], MIR [3], GSS [4], ASER [51], ER-AML [7], GDumb [48], SCR [40], OCM [21], OnPro [67] and GSA [22]) and 7 offline CL methods (DER++ [6], IL2A [75], CO 2 L [8], LUCIR [28], CCIL [45], BIC [69] and SSIL [1]).”
• Dealing with Cross-Task Class Discrimination in Online Continual Learning (2023, methodology): “Data augmentation has also been used to learn better features for CL recently [55].”
• Class-Incremental Learning: A Survey (2023, methodology): “[ 223] finds that representations with large eigenvalues transfer better and suffer less forgetting, and [221] shows that eigenvalues can be enlarged by maximizing the mutual information between old and new features.”