ZS-IL: Looking Back on Learned Experiences For Zero-Shot Incremental Learning

Abstract

Classical deep neural networks are limited in their ability to learn from emerging streams of training data. When trained sequentially on new or evolving tasks, their performance degrades sharply, making them inappropriate in real-world use cases. Existing methods tackle it by either storing old data samples or only updating a parameter set of DNNs, which, however, demands a large memory budget or spoils the flexibility of models to learn the incremented class distribution. In this paper, we shed light on an on-call transfer set to provide past experiences whenever a new class arises in the data stream. In particular, we propose a Zero-Shot Incremental Learning not only to replay past experiences the model has learned but also to perform this in a zero-shot manner. Towards this end, we introduced a memory recovery paradigm in which we query the network to synthesize past exemplars whenever a new task (class) emerges. Thus, our method needs no fixed-sized memory, besides calls the proposed memory recovery paradigm to provide past exemplars, named a transfer set in order to mitigate catastrophically forgetting the former classes. Moreover, in contrast with recently proposed methods, the suggested paradigm does not desire a parallel architecture since it only relies on the learner network. Compared to the state-of-the-art data techniques without buffering past data samples, ZS-IL demonstrates significantly better performance on the well-known datasets (CIFAR-10, Tiny-ImageNet) in both Task-IL and Class-IL settings.

References in corpus (19)

• Distilling the Knowledge in a Neural Network
• Overcoming catastrophic forgetting in neural networks
• Continual Learning Through Synaptic Intelligence
• Gradient Episodic Memory for Continual Learning
• Experience Replay for Continual Learning
• Variational Continual Learning
• Alleviating catastrophic forgetting using context-dependent gating and synaptic stabilization
• Model compression via distillation and quantization
• Continual Learning with Deep Generative Replay
• Reinforced Continual Learning
• Lifelong Learning with Dynamically Expandable Networks
• Towards Robust Evaluations of Continual Learning
• Learning to Learn without Forgetting by Maximizing Transfer and Minimizing Interference
• Dark Experience for General Continual Learning: a Strong, Simple Baseline
• Zero-Shot Knowledge Distillation in Deep Networks
• Gradient based sample selection for online continual learning
• Memory Replay GANs: learning to generate images from new categories without forgetting
• Pseudo-Recursal: Solving the Catastrophic Forgetting Problem in Deep Neural Networks
• Measuring and regularizing networks in function space