New methods for inferring fractional Brownian motion parameters using ANNs

Researchers have developed new methods for inferring parameters of fractional Brownian motion (fBM), a model used to describe anomalous diffusion. The study proposes using maximum likelihood estimation (MLE) for density measurements and an artificial neural network (ANN) trained with a squared error loss to approximate the posterior mean of the Hurst exponent ($H$). This ANN-based approach, when paired with MLE for the diffusion coefficient ($D_H$), demonstrates the lowest root mean squared error for both parameters. Analysis of vesicle trajectories from MRC5 fibroblast and HeLa cells revealed that intracellular cargoes switch between subdiffusion and superdiffusion over time, rather than maintaining fixed motion characteristics. AI

IMPACT This research introduces advanced neural network techniques for analyzing complex diffusion processes, potentially impacting fields that model particle or cellular movement.

RANK_REASON The cluster contains an academic paper detailing novel methods for parameter inference in a specific scientific model. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv stat.ML →

• Artificial neural network
• Daniel Seungmin Han
• Fractional Brownian motion
• HeLa
• Hurst exponent
• Maximum likelihood estimation
• MRC5 fibroblast

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →