NeuralFluid: Nueral Fluidic System Design and Control with Differentiable Simulation

Authors: Yifei Li, Yuchen Sun, Pingchuan Ma, Eftychios Sifakis, Tao Du, Bo Zhu, Wojciech Matusik

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental By seamlessly incorporating our differentiable fluid simulator into a learning framework, we demonstrate successful design, control, and learning results that surpass gradient-free solutions in these benchmark tasks.
Researcher Affiliation Academia Yifei Li MIT CSAIL Yuchen Sun Georgia Institute of Technology Pingchuan Ma MIT CSAIL Eftychios Sifakis University of Wisconsin-Madison Tao Du Tsinghua University, Shanghai Qi Zhi Institute Bo Zhu Georgia Institute of Technology Wojciech Matusik MIT CSAIL
Pseudocode No The paper describes the numerical simulation steps in text and mathematical equations (e.g., in Section 2.3 and Appendix B) but does not include any clearly labeled pseudocode or algorithm blocks.
Open Source Code No We plan to release our code and documentation upon acceptance.
Open Datasets No The paper introduces 'a benchmark of design, control, and learning tasks on high-fidelity, high-resolution dynamic fluid environments' and 'a suite of Gym-like environments', which are custom setups for their experiments rather than publicly available datasets with specific access information or citations.
Dataset Splits No The paper does not specify explicit training, validation, or test dataset splits (e.g., percentages or sample counts) for the environments and tasks described, nor does it mention cross-validation setups.
Hardware Specification Yes The experiment runs on a workstation with an NVIDIA RTX A6000 GPU
Software Dependencies No The paper mentions software like C++, CUDA, pybind11 [13], and PyTorch [14] but does not provide specific version numbers for these software components or libraries.
Experiment Setup Yes Initial conditions are set for all optimizations using randomly sampled values. We use Adam as our optimizer. We summarize the simulation configuration and optimization configuration as well as relevant statistics in Table 1.