Task-Aware Information Routing from Common Representation Space in Lifelong Learning
This addresses the problem of forgetting in AI systems for real-world deployment, offering a scalable solution that bridges gaps between existing methods, though it appears incremental as it builds on prior neuro-inspired and bottleneck techniques.
The paper tackles catastrophic forgetting in lifelong learning by proposing TAMiL, a method inspired by brain processes that uses task-attention modules and autoencoders to route task-relevant information, outperforming state-of-the-art approaches and reducing task interference.
Intelligent systems deployed in the real world suffer from catastrophic forgetting when exposed to a sequence of tasks. Humans, on the other hand, acquire, consolidate, and transfer knowledge between tasks that rarely interfere with the consolidated knowledge. Accompanied by self-regulated neurogenesis, continual learning in the brain is governed by a rich set of neurophysiological processes that harbor different types of knowledge, which are then integrated by conscious processing. Thus, inspired by the Global Workspace Theory of conscious information access in the brain, we propose TAMiL, a continual learning method that entails task-attention modules to capture task-specific information from the common representation space. We employ simple, undercomplete autoencoders to create a communication bottleneck between the common representation space and the global workspace, allowing only the task-relevant information to the global workspace, thus greatly reducing task interference. Experimental results show that our method outperforms state-of-the-art rehearsal-based and dynamic sparse approaches and bridges the gap between fixed capacity and parameter isolation approaches while being scalable. We also show that our method effectively mitigates catastrophic forgetting while being well-calibrated with reduced task-recency bias.