Divide and Conquer Self-Supervised Learning for High-Content Imaging
This addresses a limitation in self-supervised learning for high-content imaging in science and engineering, where complex features are critical for discovery and analysis, offering a tool to uncover overlooked features.
The paper tackles the problem of self-supervised learning methods failing to learn subtle or complex features due to dominance by simpler patterns, introducing SpliCER, a novel architecture that splits images into sections to guide learning of complex features, achieving significant downstream performance improvements in medical and geospatial imaging.
Self-supervised representation learning methods often fail to learn subtle or complex features, which can be dominated by simpler patterns which are much easier to learn. This limitation is particularly problematic in applications to science and engineering, as complex features can be critical for discovery and analysis. To address this, we introduce Split Component Embedding Registration (SpliCER), a novel architecture which splits the image into sections and distils information from each section to guide the model to learn more subtle and complex features without compromising on simpler features. SpliCER is compatible with any self-supervised loss function and can be integrated into existing methods without modification. The primary contributions of this work are as follows: i) we demonstrate that existing self-supervised methods can learn shortcut solutions when simple and complex features are both present; ii) we introduce a novel self-supervised training method, SpliCER, to overcome the limitations of existing methods, and achieve significant downstream performance improvements; iii) we demonstrate the effectiveness of SpliCER in cutting-edge medical and geospatial imaging settings. SpliCER offers a powerful new tool for representation learning, enabling models to uncover complex features which could be overlooked by other methods.