arXiv cs.AI· Shengyuan Hu, Yiyong Sun, Weihang Shan, Yongzhe Chang·· 4 小时前AI 评分31
FM-Mamba:一种带稳定性感知离散化的高效原生 2D 选择性状态空间模型
FM-Mamba: An Efficient Native 2D Selective State Space Model with Stability-Aware Discretization
AI 导读
研究者提出 FM-Mamba,一种基于 Fornasini–Marchesini(FM)模型的原生 2D 选择性状态空间模型,其耦合递推沿两个空间轴演化,并用两个独立的输入相关步长 Δ1(i,j)、Δ2(i,j) 分别控制两轴传播。
正文 · 原文
Abstract:When applied to two-dimensional features such as images, Mamba-based methods either flatten the feature map into 1D sequences and fuse multiple scans, which breaks spatial adjacency and multiplies computation, or recurse directly on the 2D grid but rescale the recurrence with a fixed input-independent factor and aggregate only the predecessors on a single anti-diagonal, leaving the discrete step unconstrained by 2D stability theory. We propose \textbf{FM-Mamba}, a native 2D selective state space model built on the general Fornasini--Marchesini (FM) model. Its coupled recurrence evolves along both spatial axes; cross-axis coupling induces a discrete diagonal-predecessor term $\tilde A_0$, and two independent input-dependent step sizes $\Delta_1(i,j)$, $\Delta_2(i,j)$ govern propagation along the two axes, so that the coefficients adapt per position and per direction. A normalized ratio $r_k=a_{0,k}/(a_{1,k}a_{2,k})$ yields a closed-form piecewise bound on the two step sizes; when a configuration falls outside the stable interval the steps are projected onto this bound, and the criterion is re-verified per position after training, which suppresses systematic local amplification without global damping. The recurrence is executed by a dual-diagonal wavefront scan: each update gathers information from two anti-diagonals, and the input enters the current state through a single projection, so that neither training nor inference flattens the feature map. Code is available at this https URL
| Subjects: | Systems and Control (eess.SY); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2611.22061 [eess.SY] |
| (or arXiv:2611.22061v1 [eess.SY] for this version) | |
| https://doi.org/10.48550/arXiv.2611.22061 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yiyong Sun Dr [view email]
[v1]
Tue, 17 Nov 2026 01:40:53 UTC (141 KB)
来源:arXiv cs.AI · arxiv.org