Synchronizing Task Behavior: Aligning Multiple Tasks during Test-Time Training
This addresses a domain-specific challenge for deploying neural networks in real-world scenarios with multiple tasks and distribution shifts, representing an incremental improvement.
The paper tackles the problem of unsynchronized task behavior in test-time training when models perform multiple tasks under domain shifts, proposing S4T to synchronize tasks and showing it outperforms state-of-the-art methods on various benchmarks.
Generalizing neural networks to unseen target domains is a significant challenge in real-world deployments. Test-time training (TTT) addresses this by using an auxiliary self-supervised task to reduce the domain gap caused by distribution shifts between the source and target. However, we find that when models are required to perform multiple tasks under domain shifts, conventional TTT methods suffer from unsynchronized task behavior, where the adaptation steps needed for optimal performance in one task may not align with the requirements of other tasks. To address this, we propose a novel TTT approach called Synchronizing Tasks for Test-time Training (S4T), which enables the concurrent handling of multiple tasks. The core idea behind S4T is that predicting task relations across domain shifts is key to synchronizing tasks during test time. To validate our approach, we apply S4T to conventional multi-task benchmarks, integrating it with traditional TTT protocols. Our empirical results show that S4T outperforms state-of-the-art TTT methods across various benchmarks.