SPNets: Differentiable Fluid Dynamics for Deep Neural Networks

TL;DR

SPNets introduces a differentiable framework integrating fluid dynamics into deep neural networks, enabling learning and control of fluid parameters and behaviors directly within the network structure.

Contribution

The paper presents SPNets, a novel neural network framework with new layers for differentiable fluid simulation, allowing end-to-end learning of fluid parameters and control policies.

Findings

Successfully learned fluid parameters from data

Performed liquid control tasks with learned policies

Demonstrated differentiable fluid dynamics within neural networks

Abstract

In this paper we introduce Smooth Particle Networks (SPNets), a framework for integrating fluid dynamics with deep networks. SPNets adds two new layers to the neural network toolbox: ConvSP and ConvSDF, which enable computing physical interactions with unordered particle sets. We use these lay- ers in combination with standard neural network layers to directly implement fluid dynamics inside a deep network, where the parameters of the network are the fluid parameters themselves (e.g., viscosity, cohesion, etc.). Because SPNets are imple- mented as a neural network, the resulting fluid dynamics are fully differentiable. We then show how this can be successfully used to learn fluid parameters from data, perform liquid control tasks, and learn policies to manipulate liquids.