Physics and Data Driven Transformer-Mamba Framework for Flow Field

Abstract

While deep learning accelerates expensive partial differential equation solving in computational fluid dynamics (CFD), existing methods like PINNs and FNOs often struggle with generalization, noise robustness, and physical consistency. We introduce the Transformer-Mamba for Flow Field (TM4FF) framework, a physics-constrained operator learning model with three key innovations: a Residual Wavelet Mamba (RWM) layer for feature denoising, a Transformer-based attention mechanism for enhanced feature fusion, and a physics-informed loss using Fourier derivatives to enforce the Navier-Stokes equations. Experiments on four CFD datasets show TM4FF achieves high accuracy and robust generalization across varying flow conditions.

The authors’ abstract, as arXiv lists it. arXiv’s listings, abstracts included, are free for anyone to reuse.

More in Machine Learning

• arXiv preprint · 24 September 2026

  Multimodal Routing and Region Refinement for Language-Guided Medical Image Segmentation

  Md Maklachur Rahman, Md Hasan Al Banna, Saraf Anjum and 2 others
• arXiv preprint · 25 September 2026

  MARCEDES: Score-based causal discovery under non-Gaussianity with continuous optimization

  Anamitra Chaudhuri, Anirban Bhattacharya and Yang Ni
• arXiv preprint · 24 September 2026

  Image Fidelity is Not Field Fidelity: Joint Thermodynamic Reconstruction and Error Localization in Neural Tomography

  Alan Hsu, Jenna Samra, Alin Razvan Paraschiv and 1 other
• arXiv preprint · 25 September 2026

  Causal Retention in Interactive Agents: Interface Factorization and Selective Adaptation

  Shengjun Zhang, Tingyi Liu, Dong Xie and 3 others
• arXiv preprint · 24 September 2026

  When Fancy Eviction Fails: Rethinking Cache Replacement For LLM Prefix Reuse

  Yiyu Liu, Minlan Yu and Juncheng Yang

All new papers in Machine Learning