MemFlow: A Lightweight Forward Memorizing Framework for Quick Domain Adaptive Feature Mapping

ArXi:2402.14598v3 Announce Type: replace-cross Deploying pretrained visual models in real-world environments often suffers from significant performance degradation due to the diversity of testing scenarios. Continuous adaptation of learning models on edge devices via unlabeled data collected from the target domain is highly effective for boosting generalization capability. However, gradient-backpropagation-based optimization of the massive parameters in deep neural networks is vastly time-consuming than forward inference, rendering online learning infeasible on low-power edge devices.