Continual Learning of a Mixed Sequence of Similar and Dissimilar Tasks

Abstract

Existing research on continual learning of a sequence of tasks focused on dealing with catastrophic forgetting, where the tasks are assumed to be dissimilar and have little shared knowledge. Some work has also been done to transfer previously learned knowledge to the new task when the tasks are similar and have shared knowledge. To the best of our knowledge, no technique has been proposed to learn a sequence of mixed similar and dissimilar tasks that can deal with forgetting and also transfer knowledge forward and backward. This paper proposes such a technique to learn both types of tasks in the same network. For dissimilar tasks, the algorithm focuses on dealing with forgetting, and for similar tasks, the algorithm focuses on selectively transferring the knowledge learned from some similar previous tasks to improve the new task learning. Additionally, the algorithm automatically detects whether a new task is similar to any previous tasks. Empirical evaluation using sequences of mixed tasks demonstrates the effectiveness of the proposed model.

References in corpus (3)

• PathNet: Evolution Channels Gradient Descent in Super Neural Networks
• Three scenarios for continual learning
• Learn to Grow: A Continual Structure Learning Framework for Overcoming Catastrophic Forgetting

Cited by in corpus (8)

• Federated Continual Learning via Knowledge Fusion: A Survey
• Continual Learning on Dynamic Graphs via Parameter Isolation
• Task Relation-aware Continual User Representation Learning
• Unified Question Generation with Continual Lifelong Learning
• Queryable Prototype Multiple Instance Learning with Vision-Language Models for Incremental Whole Slide Image Classification
• Preventing Catastrophic Forgetting in Continual Learning of New Natural Language Tasks
• Formalizing the Generalization-Forgetting Trade-off in Continual Learning
• Graceful Degradation and Related Fields