Deep Lifelong Learning Optimization Algorithm in Dense Region Fusion

Deep lifelong learning models can learn new information continuously while minimizing the impact on previously acquired knowledge, and thus adapt to changing data. However, existing optimization approaches for deep cannot simultaneously satisfy following conditions: unrestricted of data, no use old increase in model parameters. To address this problem, a algorithm based dense region fusion (DLLO-DRF) is proposed. This first obtains each stage learning, divides parameters into multiple regions parameter values. Then, dispersion distribution, are dynamically obtained from divided regions, averaged fused optimize model. Finally, extensive experiments conducted self-labeled transmission line defect dataset, results show that DLLO-DRF has best performance among various comparative algorithms.