Bidirectional Mamba state-space model for anomalous diffusion

TL;DR

This paper introduces Bi-Mamba, a bidirectional state-space deep learning model that efficiently characterizes anomalous diffusion from short trajectories, demonstrating promising results on benchmark datasets.

Contribution

The paper presents a novel bidirectional state-space deep learning architecture, Bi-Mamba, tailored for regression tasks in anomalous diffusion characterization.

Findings

Bi-Mamba accurately infers diffusion coefficients and anomalous exponents.

The model performs well on the AnDi-2 challenge datasets.

Bi-Mamba shows potential for practical applications in complex stochastic systems.

Abstract

Characterizing anomalous diffusion is crucial in order to understand the evolution of complex stochastic systems, from molecular interactions to cellular dynamics. In this work, we characterize the performances regarding such a task of Bi-Mamba, a novel state-space deep-learning architecture articulated with a bidirectional scan mechanism. Our implementation is tested on the AnDi-2 challenge datasets among others. Designed for regression tasks, the Bi-Mamba architecture infers efficiently the effective diffusion coefficient and anomalous exponent from single, short trajectories. As such, our results indicate the potential practical use of the Bi-Mamba architecture for anomalousdiffusion characterization.