Simple Lifelong Learning Machines
This work addresses the challenge of lifelong learning for AI systems by proposing a method that enhances transfer in both directions, offering a more comprehensive solution compared to incremental approaches focused solely on forgetting.
The paper tackles the problem of lifelong learning by aiming to improve performance on both future and past tasks, rather than just avoiding forgetting. It shows that a simple representation ensembling approach achieves forward and backward transfer across various datasets, outperforming reference algorithms that often fail in one or both directions.
In lifelong learning, data are used to improve performance not only on the present task, but also on past and future (unencountered) tasks. While typical transfer learning algorithms can improve performance on future tasks, their performance on prior tasks degrades upon learning new tasks (called forgetting). Many recent approaches for continual or lifelong learning have attempted to maintain performance on old tasks given new tasks. But striving to avoid forgetting sets the goal unnecessarily low. The goal of lifelong learning should be to use data to improve performance on both future tasks (forward transfer) and past tasks (backward transfer). In this paper, we show that a simple approach -- representation ensembling -- demonstrates both forward and backward transfer in a variety of simulated and benchmark data scenarios, including tabular, vision (CIFAR-100, 5-dataset, Split Mini-Imagenet, and Food1k), and speech (spoken digit), in contrast to various reference algorithms, which typically failed to transfer either forward or backward, or both. Moreover, our proposed approach can flexibly operate with or without a computational budget.