Online Continual Learning on Class Incremental Blurry Task Configuration with Anytime Inference

By Hyun-woo Koh, Jung-Woo Ha, Jonghyun Choi. 2021

89 citations recorded by the source corpus.

Despite rapid advances in continual learning, a large body of research is devoted to improving performance in the existing setups. While a handful of work do propose new continual learning setups, they still lack practicality in certain aspects. For better practicality, we first propose a novel continual learning setup that is online, task-free, class-incremental, of blurry task boundaries and subject to inference queries at any moment. We additionally propose a new metric to better measure the performance of the continual learning methods subject to inference queries at any moment. To address the challenging setup and evaluation protocol, we propose an effective method that employs a new memory management scheme and novel learning techniques. Our empirical validation demonstrates that the proposed method outperforms prior arts by large margins. Code and data splits are available at https://github.com/naver-ai/i-Blurry.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, 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. Ziwei Liu
5. Jun Zhu
6. Hang Su
7. Liyuan Wang
8. Qiwen Wang

In Computer Vision and Pattern Recognition

1. Han-Jia Ye
2. De-chuan Zhan
3. Jun Zhu
4. Ziwei Liu
5. Luc Van Gool
6. Da-Wei Zhou
7. Hang Su
8. Bernt Schiele

In Meta & Continual Learning

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

Citations in the corpus

• A Comprehensive Survey of Continual Learning: Theory, Method and Application (2023, background): “entropy (ER-ACE [47]) or constrained gradient directions (GEM [277], A-GEM [65]) of the old and new training samples, repeated rehearsal with data augmentation (RAR [498]), properly adjusting the learning rate (PoLRS [50], CLIB [224]), training from scratch with the memory buffer”
• Deep Class-Incremental Learning: A Survey (2023, background): “When Yb ∩ Yb′ 6= ∅, the setting is called blurry class-incremental learning (Blurry CIL) [34], [35], [36].”
• One Size Fits All for Semantic Shifts: Adaptive Prompt Tuning for Continual Learning (2023, background): “These methods can be further categorized into universal and specific prompting methods, depending on how the prompts are managed for sequentially arriving tasks.”
• Coherent Temporal Synthesis for Incremental Action Segmentation (2024, unknown): “The performance of an incremental learning model is evaluated over all seen classes, i.e ., ), but when class overlap occurs, the task is referred to as blurry class-incremental learning (Blurry CIL) [4,5,22].”
• SDDGR: Stable Diffusion-Based Deep Generative Replay for Class Incremental Object Detection (2024, methodology): “Among these, Replay methods are most frequently utilized for their simple yet powerful effects and can be categorized into two types: partial experience replay (ER) [2, 4, 14, 27, 44, 47] and generative replay (GR) [6, 12, 23, 58, 66, 67].”
• Online Continual Learning For Interactive Instruction Following Agents (2024, methodology): “We observe that our CAMA outperforms all rehearsal-based approaches ( i.e. , ER, MIR, and CLIB) for all metrics in both seen and unseen environments in both Behavior-IL and Environment-IL setups.”
• Continual Learning on a Diet: Learning from Sparsely Labeled Streams Under Constrained Computation (2024, background): “Some work considers releasing the task boundary constraint and explores algorithms with unknown boundaries (Aljundi et al., 2019) or blurry task boundaries (Koh et al., 2021).”