Search papers, labs, and topics across Lattice.
This paper introduces MoNo, a Multiscale Optimal Transport Neural Operator designed to enhance the solving of Partial Differential Equations (PDEs) on general geometries by addressing the limitations of existing projection mechanisms that lead to assignment imbalances among latent tokens. By employing CoTAP, a novel method that formulates cross-space assignments as an entropy-regularized optimal transport problem, MoNo achieves balanced projections and stable latent spaces, which are crucial for effective hierarchical architectures. Experimental results show that MoNo significantly outperforms current state-of-the-art neural operators in terms of both prediction accuracy and computational efficiency.
MoNo's innovative approach to optimal transport ensures stable latent spaces, preventing token collapse and enabling efficient learning of long-range physical interactions in PDEs.
Transformer-based neural operators have achieved substantial progress in solving Partial Differential Equations (PDEs) by projecting spatial observations into compact latent tokens and learning physical interactions in latent spaces. However, we reveal that existing learnable projection mechanisms cannot ensure stable and balanced assignments from observation points to latent tokens, causing some latent tokens to be over-assigned while others remain underutilized. This limitation further restricts the design of hierarchical architectures, as assignment imbalance is continuously inherited and amplified across latent spaces, eventually causing severe token collapse in deeper spaces. To address these issues, we propose MoNo (Multiscale Optimal Transport Neural Operator), a progressive multiscale neural operator that efficiently solves PDEs on general geometries through stable latent-space construction. At its core is CoTAP (Cross-scale Optimal Transport Assignment and Projection), a novel latent-space construction method that formulates cross-space assignment between adjacent spaces as an entropy-regularized optimal transport problem, thereby constructing balanced bidirectional projections and stable latent spaces. CoTAP also ensures stable information transfer across multiple latent spaces, further enabling multiscale architectures on general geometries, which in turn support more efficient learning of long-range physical interactions. Extensive experiments demonstrate that MoNo outperforms existing state-of-the-art neural operators in both prediction performance and computational efficiency. Code is available at https://github.com/ZijiangY1116/MoNo.