Contextual Transformation Networks for Online Continual Learning

By Quang Pham, Chenghao Liu, Doyen Sahoo, Steven C. H. Hoi. 2021

61 citations recorded by the source corpus.

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

Citations in the corpus

Incorporating neuro-inspired adaptability for continual learning in artificial intelligence (2023, methodology): “The CORe50 dataset [63] includes 50 handheld objects with smoothly-changed observations of size 128 × 128, randomly split into 10 tasks [73].”
Improving Task-free Continual Learning by Distributionally Robust Memory Evolution (2022, methodology): “…& Ranzato, 2017; Nguyen et al., 2018; Kirkpatrick et al., 2017; Zenke et al., 2017; von Oswald et al., 2019; Wang et al., 2021; Saha et al., 2021; Pham et al., 2021; Wang et al., 2022) are only applicable to the task-aware setting, where there are clear task deﬁnitions and boundaries among the…”
DualPrompt: Complementary Prompting for Rehearsal-free Continual Learning (2022, methodology): “Some other well-known methods [17, 11, 44] are not compared here because they either have been outperformed by our compared methods or only verified on simpler task-incremental setting.”
Continual Normalization: Rethinking Batch Normalization for Online Continual Learning (2022, background): “…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; Aljundi et al., 2018; Ritter et al., 2018), to fast and slow frameworks (Pham et al., 2020; 2021).”
Learning to Prompt for Continual Learning (2021, background): “Lastly, recent work of CTN [45] and DualNet [44] start to consider knowledge management via a controller that models task-level information in addition to a back-bone model.”
Else-Net: Elastic Semantic Network for Continual Action Recognition from Skeleton Data (2021, methodology): “To evaluate the abilities of the proposed Else-Net in both effectively learning the new actions and mitigating catastrophic forgetting, we follow the metrics introduced in [25] for continual learning performance evaluation.”
Mitigating Forgetting in Online Continual Learning with Neuron Calibration (2021, methodology): “Our work is mostly related to [25], which engages task-speciﬁc embeddings to calibrate the feature output of neural networks and train the task-speciﬁc parameters with hold-out validation data to promote generalization.”
How to Reuse and Compose Knowledge for a Lifetime of Tasks: A Survey on Continual Learning and Functional Composition (2022, methodology): “As another example, Pham et al. (2021b) used a dual memory to train a set of shared neural network layers and a task-specific controller to transform the features of the shared layers.”
DualNet: Continual Learning, Fast and Slow Supplementary Materials (2021, methodology): “…in our experiments in Table 1. miniImageNet [11] 100 50,000 10,000 3 × 84 × 84 3 × 128 × 128 CORe50 [5] 50 119,894 44,971 3 × 84 × 84 3 × 128 × 128 We consider three standard metrics to evaluate the methods [8]: ACC [6] (higher is better), FM [2] (lower is better), and LA [10] (higher is better).”
Experts Collaboration Learning for Continual Multi-Modal Reasoning (2023, methodology): “classification [49], [50], [51] and object detection [52], [53].”