Few-Shot Self Reminder to Overcome Catastrophic Forgetting

By Junfeng Wen, Yanshuai Cao, Ruitong Huang. 2018

24 citations recorded by the source corpus.

Deep neural networks are known to suffer the catastrophic forgetting problem, where they tend to forget the knowledge from the previous tasks when sequentially learning new tasks. Such failure hinders the application of deep learning based vision system in continual learning settings. In this work, we present a simple yet surprisingly effective way of preventing catastrophic forgetting. Our method, called Few-shot Self Reminder (FSR), regularizes the neural net from changing its learned behaviour by performing logit matching on selected samples kept in episodic memory from the old tasks. Surprisingly, this simplistic approach only requires to retrain a small amount of data in order to outperform previous methods in knowledge retention. We demonstrate the superiority of our method to the previous ones in two different continual learning settings on popular benchmarks, as well as a new continual learning problem where tasks are designed to be more dissimilar.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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 Computer Vision and Pattern Recognition

Citations in the corpus

Linear Mode Connectivity in Multitask and Continual Learning (2020, background): “In this sense it connects to topics such as MetaLearning (Beaulieu et al., 2020; Jerfel et al., 2019; He & Jaeger, 2018; Riemer et al., 2018), Few-Shot Learning (Wen et al., 2018; Gidaris & Komodakis, 2018), Multi-task and Transfer Learning (He & Jaeger, 2018; Jerfel et al., 2019).”
Understanding the Role of Training Regimes in Continual Learning (2020, methodology): “Continual Learning as an emerging ﬁeld in Artiﬁcial Intelligence is connected to many other areas such as Meta Learning [4, 33, 22, 59], Few Shot Learning [71, 19], Multi-task and Transfer Learning [22, 33] and the closely related problem of exploring task boundary or task detection [57, 2, 23, 34].”
Orthogonal Gradient Descent for Continual Learning (2019, background): “Few shot learning also serves the same purpose and can be leveraged in continual learning (Wen et al., 2018; Gi-daris and Komodakis, 2018) and vice versa.”
Variational Prototype Replays for Continual Learning (2019, background): “Moreover, to synthesize the historical data reasonably well, the size of the generative model is usually very large and expensive in terms of memory resources (Wen et al., 2018).”
Prototype Reminding for Continual Learning (2019, methodology): “Specifically, in FSR, there are two variants: FSR-KLD for logits matching via Kullback–Leibler Divergence loss and FSR-MSE for logits distillation via L2 distance loss.”