Meta-Learning with Sparse Experience Replay for Lifelong Language Learning

Abstract

Lifelong learning requires models that can continuously learn from sequential streams of data without suffering catastrophic forgetting due to shifts in data distributions. Deep learning models have thrived in the non-sequential learning paradigm; however, when used to learn a sequence of tasks, they fail to retain past knowledge and learn incrementally. We propose a novel approach to lifelong learning of language tasks based on meta-learning with sparse experience replay that directly optimizes to prevent forgetting. We show that under the realistic setting of performing a single pass on a stream of tasks and without any task identifiers, our method obtains state-of-the-art results on lifelong text classification and relation extraction. We analyze the effectiveness of our approach and further demonstrate its low computational and space complexity.

References in corpus (18)

• Overcoming catastrophic forgetting in neural networks
• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• Theoretical Models of Learning to Learn
• Riemannian Walk for Incremental Learning: Understanding Forgetting and Intransigence
• Continual Learning Through Synaptic Intelligence
• PathNet: Evolution Channels Gradient Descent in Super Neural Networks
• Efficient Lifelong Learning with A-GEM
• On First-Order Meta-Learning Algorithms
• The Natural Language Decathlon: Multitask Learning as Question Answering
• FearNet: Brain-Inspired Model for Incremental Learning
• Towards Robust Evaluations of Continual Learning
• Learning and Evaluating General Linguistic Intelligence
• LAMOL: LAnguage MOdeling for Lifelong Language Learning
• An Empirical Study of Example Forgetting during Deep Neural Network Learning
• Learning to Continually Learn
• AI-GAs: AI-generating algorithms, an alternate paradigm for producing general artificial intelligence
• Backpropamine: training self-modifying neural networks with differentiable neuromodulated plasticity
• Optimal Continual Learning has Perfect Memory and is NP-hard

Cited by in corpus (4)

• Prototype-Guided Memory Replay for Continual Learning
• DRILL: Dynamic Representations for Imbalanced Lifelong Learning
• Continual Learning for Text Classification with Information Disentanglement Based Regularization
• Meta-Learning with Variational Semantic Memory for Word Sense Disambiguation