BiRT：视觉Transformer中用于持续学习的仿生回放
BiRT: Bio-inspired Replay in Vision Transformers for Continual Learning
AI总结：
针对视觉Transformer持续学习中表征回放多样性不足导致的过拟合问题，提出BiRT方法，通过在各阶段引入建设性噪声并强制与指数移动平均预测一致，在多个基准上取得优于原始图像回放和朴素表征回放的性能，且内存高效、鲁棒。
摘要：
深度神经网络持续学习并适应一系列任务的能力，由于对先前学习任务的灾难性遗忘而一直面临挑战。另一方面，人类在终身学习过程中具有显著的能力，能够获取、同化并跨任务迁移知识，而不会发生灾难性遗忘。大脑的这种多功能性可归因于通过互补学习系统对抽象经验进行排练（回放）。然而，视觉Transformer中的表征回放缺乏多样性，导致过拟合，进而与原始图像回放相比性能显著下降。因此，我们提出BiRT，一种基于表征回放的新型持续学习方法，使用视觉Transformer。具体而言，我们在视觉Transformer的各个阶段引入建设性噪声，并针对工作模型的指数移动平均强制预测一致性。我们的方法在多个具有挑战性的持续学习基准上，相较于原始图像回放和朴素表征回放，提供了一致的性能提升，同时具有内存效率，并且对自然和对抗性损坏具有鲁棒性。
英文摘要：
The ability of deep neural networks to continually learn and adapt to a sequence of tasks has remained challenging due to catastrophic forgetting of previously learned tasks. Humans, on the other hand, have a remarkable ability to acquire, assimilate, and transfer knowledge across tasks throughout their lifetime without catastrophic forgetting. The versatility of the brain can be attributed to the rehearsal of abstract experiences through a complementary learning system. However, representation rehearsal in vision transformers lacks diversity, resulting in overfitting and consequently, performance drops significantly compared to raw image rehearsal. Therefore, we propose BiRT, a novel representation rehearsal-based continual learning approach using vision transformers. Specifically, we introduce constructive noises at various stages of the vision transformer and enforce consistency in predictions with respect to an exponential moving average of the working model. Our method provides consistent performance gain over raw image and vanilla representation rehearsal on several challenging CL benchmarks, while being memory efficient and robust to natural and adversarial corruptions.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
