Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Learning Similarity Metrics for Volumetric Simulations with Multiscale CNNs
Authors: Georg Kohl, Li-Wei Chen, Nils Thuerey
AAAI 2023 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | All metrics were evaluated on the volumetric data from Sec. 4, which contain a wide range of test sets that differ strongly from the training data. |
| Researcher Affiliation | Academia | Georg Kohl, Li-Wei Chen, Nils Thuerey Technical University of Munich EMAIL |
| Pseudocode | No | The paper describes methods and a network architecture using figures and text but does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | Our source code, datasets, and ready-to-use models are available at https://github.com/tum-pbs/VOLSIM. |
| Open Datasets | Yes | Our source code, datasets, and ready-to-use models are available at https://github.com/tum-pbs/VOLSIM. Furthermore, we use adjusted versions of the noise integration for two test sets, by adding noise to the density instead of the velocity in the Advection-Diffusion model (Adv D) and overlaying background noise in the liquid simulation (Liq N). We create seven test sets via method [B]. Four come from the Johns Hopkins Turbulence Database JHTDB (Perlman et al. 2007) that contains a large amount of direct numerical simulation (DNS) data... One additional test set (SF) via temporal translations is based on Scalar Flow (Eckert, Um, and Thuerey 2019)... |
| Dataset Splits | Yes | The corresponding validation sets are generated with a separate set of random seeds. |
| Hardware Specification | No | The paper discusses memory limitations for training but does not provide specific details about the hardware (e.g., GPU/CPU models, memory specifications) used for running experiments. |
| Software Dependencies | No | The paper mentions using the 'Adam optimizer' but does not provide specific version numbers for any software dependencies, libraries, or programming languages used in the experiments. |
| Experiment Setup | Yes | The final metric model was trained with the Adam optimizer with a learning rate of 10^-4 for 30 epochs via early stopping. |