Instance-Level Meta Normalization
This addresses the normalization challenge in deep learning, particularly for scenarios with small mini-batches, but it is incremental as it builds upon existing instance-level normalization schemes.
The paper tackles the problem of learning-to-normalize by proposing Instance-Level Meta Normalization (ILM~Norm), which predicts normalization parameters through feature feed-forward and gradient back-propagation paths, resulting in consistent performance improvements across various network architectures and tasks.
This paper presents a normalization mechanism called Instance-Level Meta Normalization (ILM~Norm) to address a learning-to-normalize problem. ILM~Norm learns to predict the normalization parameters via both the feature feed-forward and the gradient back-propagation paths. ILM~Norm provides a meta normalization mechanism and has several good properties. It can be easily plugged into existing instance-level normalization schemes such as Instance Normalization, Layer Normalization, or Group Normalization. ILM~Norm normalizes each instance individually and therefore maintains high performance even when small mini-batch is used. The experimental results show that ILM~Norm well adapts to different network architectures and tasks, and it consistently improves the performance of the original models. The code is available at url{https://github.com/Gasoonjia/ILM-Norm.