MixStyle Neural Networks for Domain Generalization and Adaptation
This addresses the longstanding issue of domain generalization in machine learning, enabling better performance on unseen domains without extra data or model capacity, though it is an incremental improvement based on style transfer insights.
The paper tackles the problem of neural networks failing to generalize to unseen data with domain shifts by proposing MixStyle, a plug-and-play module that mixes feature statistics during training to synthesize new domains, resulting in significant boosts in out-of-distribution generalization across tasks like image recognition and reinforcement learning.
Neural networks do not generalize well to unseen data with domain shifts -- a longstanding problem in machine learning and AI. To overcome the problem, we propose MixStyle, a simple plug-and-play, parameter-free module that can improve domain generalization performance without the need to collect more data or increase model capacity. The design of MixStyle is simple: it mixes the feature statistics of two random instances in a single forward pass during training. The idea is grounded by the finding from recent style transfer research that feature statistics capture image style information, which essentially defines visual domains. Therefore, mixing feature statistics can be seen as an efficient way to synthesize new domains in the feature space, thus achieving data augmentation. MixStyle is easy to implement with a few lines of code, does not require modification to training objectives, and can fit a variety of learning paradigms including supervised domain generalization, semi-supervised domain generalization, and unsupervised domain adaptation. Our experiments show that MixStyle can significantly boost out-of-distribution generalization performance across a wide range of tasks including image recognition, instance retrieval and reinforcement learning.