A polar prediction model for learning to represent visual transformations
This work addresses the challenge of representing visual transformations for organisms to make accurate temporal predictions, with potential applications in neuroscience and computer vision, though it appears incremental as it builds on existing self-supervised and group-theoretic methods.
The authors tackled the problem of learning visual representations for temporal prediction by proposing a self-supervised framework based on polar architecture, which achieved better prediction performance than traditional motion compensation and rivaled conventional deep networks while maintaining interpretability and speed.
All organisms make temporal predictions, and their evolutionary fitness level depends on the accuracy of these predictions. In the context of visual perception, the motions of both the observer and objects in the scene structure the dynamics of sensory signals, allowing for partial prediction of future signals based on past ones. Here, we propose a self-supervised representation-learning framework that extracts and exploits the regularities of natural videos to compute accurate predictions. We motivate the polar architecture by appealing to the Fourier shift theorem and its group-theoretic generalization, and we optimize its parameters on next-frame prediction. Through controlled experiments, we demonstrate that this approach can discover the representation of simple transformation groups acting in data. When trained on natural video datasets, our framework achieves better prediction performance than traditional motion compensation and rivals conventional deep networks, while maintaining interpretability and speed. Furthermore, the polar computations can be restructured into components resembling normalized simple and direction-selective complex cell models of primate V1 neurons. Thus, polar prediction offers a principled framework for understanding how the visual system represents sensory inputs in a form that simplifies temporal prediction.