General Incremental Learning with Domain-aware Categorical Representations

By Jiangwei Xie, Shipeng Yan, Xuming He. 2022

50 citations recorded by the source corpus.

Continual learning is an important problem for achieving human-level intelligence in real-world applications as an agent must continuously accumulate knowledge in response to streaming data/tasks. In this work, we consider a general and yet under-explored incremental learning problem in which both the class distribution and class-specific domain distribution change over time. In addition to the typical challenges in class incremental learning, this setting also faces the intra-class stability-plasticity dilemma and intra-class domain imbalance problems. To address above issues, we develop a novel domain-aware continual learning method based on the EM framework. Specifically, we introduce a flexible class representation based on the von Mises-Fisher mixture model to capture the intra-class structure, using an expansion-and- reduction strategy to dynamically increase the number of components according to the class complexity. Moreover, we design a bi-level balanced memory to cope with data imbalances within and across classes, which combines with a distillation loss to achieve better inter- and intra-class stability-plasticity trade-off. We conduct exhaustive experiments on three benchmarks: iDigits, iDomainNet and iCIFAR-20. The results show that our approach consistently outperforms previous methods by a significant margin, demonstrating its superiority.

Subfields: Artificial Intelligence, Representation & Self-Supervised Learning, 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 Artificial Intelligence

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

In Representation & Self-Supervised Learning

1. Han-Jia Ye
2. De-chuan Zhan
3. Da-Wei Zhou
4. Liyuan Wang
5. Lanqing Hong
6. Jun Zhu
7. Xingxing Zhang
8. Xiaopeng Hong

In Meta & Continual Learning

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

Citations in the corpus

• A Comprehensive Survey of Continual Learning: Theory, Method and Application (2023, background): “Some recent work has focused on combinatorial scenarios [18], [40], [41], [42], [43] in order to simulate real-world complexity.”
• The Ideal Continual Learner: An Agent That Never Forgets (2023, background): “, 2022), or combinations thereof (Yan et al., 2021; Xie et al., 2022; Wang et al., 2022b; Frascaroli et al., 2023); for surveys, see, e.”
• Generative Negative Text Replay for Continual Vision-Language Pretraining (2022, background): “Continual Learning Continual learning [54,13,43,48] aims to integrate novel knowledge in a sequential fashion where old data are usually unavailable.”
• S-Prompts Learning with Pre-trained Transformers: An Occam's Razor for Domain Incremental Learning (2022, methodology): “Recently, some of work like [48, 31, 29, 36, 66, 18, 78] adapts the CIL methods to the DIL problems such as continual deepfake detection, where each domain distinguishes between deepfakes and reals.”
• Deep Class-Incremental Learning: A Survey (2023, background): “[33] proposes a unified framework to handle CIL with concept drift.”
• Class-Incremental Learning: A Survey (2023, methodology): “Specifically, RMM is a specific technique to efficiently organize the memory budget, which can be orthogonally combined with other methods, and we combine it with FOSTER, denoted as FOSTER+RMM. Similarly, we combine AANets with LUCIR [89], denoted as LUCIR+AANets.”