Machine Learning with Physics Knowledge for Prediction: A Survey
Authors: Joe Watson, Chen Song, Oliver Weeger, Theo Gruner, An Thai Le, Kay Hansel, Ahmed Hendawy, Oleg Arenz, Will Trojak, Miles Cranmer, Carlo D'Eramo, Fabian Buelow, Tanmay Goyal, Jan Peters, Martin W Hoffmann
TMLR 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | This survey examines the broad suite of methods and models for combining machine learning with physics knowledge for prediction and forecasting, with a focus on partial differential equations. These methods have attracted significant interest because of their potential impact on the advancement of scientific research and industrial practices, promising improvements to using smallor large-scale datasets and expressive predictive models with useful inductive biases. |
| Researcher Affiliation | Collaboration | 1Department of Computer Science, Technical University of Darmstadt, Germany 2Systems AI for Robot Learning, German Research Center for AI (DFKI), Germany 3Hessian Center for Artificial Intelligence (hessian.AI), Germany 4Center for Artificial Intelligence and Data Science, University of Würzburg, Germany 5Centre for Cognitive Science, Technical University of Darmstadt, Germany 6ABB Corporate Research Center, Mannheim, Germany 7Department of Mechanical Engineering, Technical University of Darmstadt, Germany 8IBM Research UKI, United Kingdom 9Data Intensive Science, University of Cambridge, United Kingdon Now at the Oxford Robotics Institute, University of Oxford |
| Pseudocode | Yes | Algorithm 1 Meta learning with MAML and Reptile |
| Open Source Code | No | The paper is a survey and discusses various open-source libraries and projects (e.g., Deep XDE, neuromancer, NVIDIA/modulus, Sci ML, neuraloperator/neuraloperator, Neuro Diff Gym/neurodiffeq) developed by others in Appendix A. However, it does not explicitly state that the authors of this survey paper are releasing their own code for the work described in this paper. |
| Open Datasets | No | The paper is a survey and references various datasets used by other works in the field (e.g., weather data) and benchmark suites (PDEBench). However, it does not use or provide access information for a specific dataset that was used to conduct its own empirical studies or experiments. It discusses datasets as part of the broader field it surveys. |
| Dataset Splits | No | The paper is a survey and does not conduct its own empirical experiments. Therefore, it does not provide details about dataset splits for training, validation, or testing. |
| Hardware Specification | No | The paper is a survey and does not describe any specific hardware used for its own work. It discusses the use of hardware like GPUs by other researchers in the field in general terms but does not specify hardware for any experiments conducted within this paper. |
| Software Dependencies | No | The paper is a survey and, while it discusses many open-source software libraries and frameworks in Appendix A, it does not list specific version numbers for software dependencies used to conduct its own research or experiments. The mentioned libraries are part of the ecosystem being surveyed, not dependencies for the survey itself. |
| Experiment Setup | No | The paper is a survey and does not describe any specific experimental setup, hyperparameters, or training configurations as it does not present its own empirical experiments. |