Neural Experts: Mixture of Experts for Implicit Neural Representations
This work addresses the challenge of global constraints in INRs for researchers in fields like computer vision and signal processing, offering an incremental improvement over existing methods.
The paper tackles the problem of learning implicit neural representations (INRs) by proposing a mixture of experts (MoE) approach that subdivides the domain and fits locally, resulting in improved speed, accuracy, and memory requirements across reconstruction tasks.
Implicit neural representations (INRs) have proven effective in various tasks including image, shape, audio, and video reconstruction. These INRs typically learn the implicit field from sampled input points. This is often done using a single network for the entire domain, imposing many global constraints on a single function. In this paper, we propose a mixture of experts (MoE) implicit neural representation approach that enables learning local piece-wise continuous functions that simultaneously learns to subdivide the domain and fit locally. We show that incorporating a mixture of experts architecture into existing INR formulations provides a boost in speed, accuracy, and memory requirements. Additionally, we introduce novel conditioning and pretraining methods for the gating network that improves convergence to the desired solution. We evaluate the effectiveness of our approach on multiple reconstruction tasks, including surface reconstruction, image reconstruction, and audio signal reconstruction and show improved performance compared to non-MoE methods.