EEML: Ensemble Embedded Meta-Learning

AbstractTo accelerate learning process with few samples, meta-learning resorts to prior knowledge from previous tasks. However, the inconsistent task distribution and heterogeneity is hard be handled through a global sharing model initialization. In this paper, based on gradient-based meta-learning, we propose an ensemble embedded algorithm (EEML) that explicitly utilizes multi-model-ensemble organize into diverse specific experts. We rely embedding cluster mechanism deliver tasks matching experts in training instruct how collaborate test phase. As result, multi can focus their own area of expertise cooperate upcoming solve heterogeneity. The experimental results show proposed method outperforms recent state-of-the-arts easily few-shot problem, which validates importance differentiation cooperation.KeywordsMeta-learningEnsemble-learningFew-shot