Understanding nanoparticle motion in complex liquids via Liquid phase transmission electron microscopy is limited by the difficulty of linking observed dynamics to underlying interaction forces without tractable physical models. The authors introduce a physics-informed transformer-based generative model that learns and reveals the stochastic, heterogeneous motion of nanoparticles, capturing key viscoelastic features of their environment.
- Zain Shabeeb
- Naisargi Goyal
- Vida Jamali
