Gallery

SFI examples gallery. Each script demonstrates a complete inference pipeline: simulate (or load) data, infer forces and diffusion, and validate the results. Use the tags below each thumbnail to filter by topic.

Getting started: end-to-end inference (Ornstein–Uhlenbeck)

Getting started: end-to-end inference (Ornstein–Uhlenbeck)

2D limit cycle — nonlinear overdamped inference

2D limit cycle — nonlinear overdamped inference

Lorenz attractor — overdamped inference

Lorenz attractor — overdamped inference

Lotka–Volterra ecosystem — sparse network recovery

Lotka–Volterra ecosystem — sparse network recovery

Experimental-data workflow template

Experimental-data workflow template

Custom basis with extras — multi-experiment traps

Custom basis with extras — multi-experiment traps

Multiplicative noise — the Landauer blowtorch

Multiplicative noise — the Landauer blowtorch

Anisotropic diffusion tensor field

Anisotropic diffusion tensor field

Van der Pol oscillator — underdamped inference

Van der Pol oscillator — underdamped inference

Velocity-dependent noise — underdamped multiplicative diffusion

Velocity-dependent noise — underdamped multiplicative diffusion

Home ranges in a shared landscape, from noisy gappy data

Home ranges in a shared landscape, from noisy gappy data

Time-dependent forcing — protocols as extras

Time-dependent forcing — protocols as extras

Learning a time-dependent force field — time-Fourier basis

Learning a time-dependent force field — time-Fourier basis

Diagnostics — assessing fit quality

Diagnostics — assessing fit quality

Overdamped or underdamped? Classifying dynamics from data

Overdamped or underdamped? Classifying dynamics from data

Aligning active Brownian particles — generic pairs API

Aligning active Brownian particles — generic pairs API

Nonreciprocal ABPs at large scale — 3 000 particles

Nonreciprocal ABPs at large scale — 3 000 particles

Discovering Toner–Tu hydrodynamics from agent-based flocking

Discovering Toner–Tu hydrodynamics from agent-based flocking

Discovering active-nematic hydrodynamics from a bacterial swarm

Discovering active-nematic hydrodynamics from a bacterial swarm

Gray-Scott reaction-diffusion: SPDE inference

Gray-Scott reaction-diffusion: SPDE inference

Advanced

These examples push the parametric estimators further: neural-network force fields, multi-experiment fitting with shared parameters, and underdamped multi-particle systems. Start with the main gallery above if you are new to SFI; the regime table in the Running-inference guide tells you when these tools are the right choice.

Neural-network force field — Müller-Brown potential

Neural-network force field — Müller-Brown potential

Multi-experiment ABP inference

Multi-experiment ABP inference

3D flocking — underdamped multi-particle inference

3D flocking — underdamped multi-particle inference