{"success":true,"data":{"author":"Polymathic AI (The Well); Suryanarayana Maddu, Flatiron Institute CCB","bucketAccessType":"public","bucketName":"polymathic-ai/active_matter","bucketPrefix":"","category":"Physics","cloudProvider":"huggingface","createdAt":"2026-11-14T01:01:19.693Z","description":"Simulations of a continuum theory for the dynamics of N rod-like active particles immersed in a Stokes fluid, a model of biological active matter such as motor-driven filament suspensions. Each trajectory has 81 time steps of 256 by 256 images, stored every 0.25 s from 0 to 20 s on a periodic 10 by 10 domain. Fields are concentration, velocity, orientation tensor and strain-rate tensor. The datasheet lists 5 trajectories per parameter set, varying the active dipole strength alpha from -1 to -5 and the alignment parameter zeta, with beta fixed at 0.8. Listing data/ (measured 2026-11-13) shows train/ (45 files, 42.5 GB, one per parameter set), valid/ (16 files, 6.3 GB) and test/ (21 files, 7.0 GB); about 56 GB in the repository. Files are HDF5 in the common The Well format; open them with the_well package (WellDataModule with base path hf://datasets/polymathic-ai/) or directly with h5py. Part of The Well (Ohana et al., NeurIPS 2024 Datasets and Benchmarks, arXiv:2412.00568); released under CC-BY-4.0, cite The Well and the simulation's source papers.","displayTitle":"The Well: Active Matter, Rod-like Active Particles in a Stokes Fluid","id":"9acfc291-cea2-57da-8822-e2e284828fa7","license":"CC-BY-4.0","licenseEvidenceUrl":"https://arxiv.org/html/2412.00568","objectCountEstimate":88,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2308.06675","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"polymathic-ai/active_matter\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"polymathic-ai/active_matter\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=polymathic-ai, license:cc-by-4.0, downloads=3200"],"format":"mixed","generatedAt":"2026-11-14T01:01:24.722Z","layout":{"example":"data/test/active_matter_L_10.0_zeta_1.0_alpha_-1.0.hdf5","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"data/test/active_matter_L_10.0_zeta_1.0_alpha_-1.0.hdf5","truncatedAt":512}],"sampleObjects":["active_matter.yaml","data/test/active_matter_L_10.0_zeta_1.0_alpha_-1.0.hdf5","data/test/active_matter_L_10.0_zeta_1.0_alpha_-2.0.hdf5","data/test/active_matter_L_10.0_zeta_11.0_alpha_-1.0.hdf5","data/test/active_matter_L_10.0_zeta_11.0_alpha_-2.0.hdf5"],"sidecars":[{"bytes":"---\nlanguage:\n- en\nlicense: cc-by-4.0\ntags:\n- physics\n- biology\ntask_categories:\n- time-series-forecasting\n- other\ntask_ids:\n- multivariate-time-series-forecasting\n---\n\n# How To Load from HuggingFace Hub\n\n1. Be sure to have `the_well` installed (`pip install the_well`)\n2. Use the `WellDataModule` to retrieve data as follows:\n\n```python\nfrom the_well.benchmark.data import WellDataModule\n\n# The following line may take a couple of minutes to instantiate the datamodule\ndatamodule = WellDataModule(\n    \"hf://datasets/polymathic-ai/\",\n    \"active_matter_cloud_optimized\",\n)\ntrain_dataloader = datamodule.train_dataloader()\n\nfor batch in dataloader:\n    # Process training batch\n    ...\n```\n\n# Active matter simulations\n\n**One line description of the data:**  Modeling and simulation of biological active matter.\n\n**Longer description of the data:** Simulation of a continuum theory describing the dynamics of \\\\(N\\\\) rod-like active particles immersed in a Stokes fluid having linear dimension \\\\(L\\\\) and colume \\\\(L^2\\\\).\n\n**Associated paper**: [Paper](https://arxiv.org/abs/2308.06675).\n\n**Domain expert**: [Suryanarayana Maddu](https://sbalzarini-lab.org/?q=alumni/surya), Center for Computaional Biology, Flatiron Institute.\n\n**Code or software used to generate the data**: [Github repository](https://github.com/SuryanarayanaMK/Learning_closures/tree/master).\n\n**Equations**: Equations (1) to (5) of the associated paper.\n\n\n![Gif](https://users.flatironinstitute.org/~polymathic/data/the_well/datasets/active_matter/gif/concentration_notnormalized.gif)\n\n| Dataset    | FNO | TFNO  | Unet | CNextU-net\n|:-:|:-:|:-:|:-:|:-:|\n| `active_matter`  | 0.3691  | 0.3598 |0.2489|\\\\(\\mathbf{0.1034}\\\\)|\n\nTable: VRMSE metrics on test sets (lower is better). Best results are shown in bold. VRMSE is scaled such that predicting the mean value of the target field results in a score of 1.\n\n\n\n\n\n# About the data\n\n**Dimension of discretized data:** 81 time-steps of 256 \\\\(\\times\\\\) 256 images per trajectory.\n\n**Fields available in the data:** concentration (scalar field),\nvelocity (vector field), orientation tensor (tensor field), strain-rate tensor (tensor field).\n\n\n**Number of trajectories:** \\\\(5\\\\) trajectories per parameter-set, each trajectory being generated with a different initialization of the state field {\\\\(c,D,U\\\\)}.\n\n**Size of the ensemble of all simulations:** 51.3 GB.\n\n**Grid type:** Uniform grid, cartesian coordinates.\n\n**Initial conditions:** The concentration is set to constant value \\\\(c(x,t)=1\\\\) and the orientation tensor is initialized as plane-wave perturbation about the isotropic state.\n\n**Boundary conditions:** Periodic boundary conditions.\n\n**Simulation time-step:** \\\\(3.90625\\times 10^{-4}\\\\) seconds.\n\n**Data are stored separated by (\\\\(\\Delta t\\\\)):** 0.25 seconds.\n\n**Total time range (\\\\(t_{min}\\\\) to \\\\(t_{max}\\\\)):** \\\\(0\\\\) to \\\\(20\\\\) seconds.\n\n**Spatial domain size (\\\\(L_x\\\\), \\\\(L_y\\\\)):** \\\\(L_x=10\\\\) and \\\\(L_y=10\\\\).\n\n**Set of coefficients or non-dimensional parameters evaluated:** \\\\(\\alpha =\\\\) {-1,-2,-3,-4,-5}; \\\\(\\beta  =\\\\) {0.8};\n\\\\(\\zeta =\\\\) {1,3,5,7,9,11,13,15,17}.\n\n**Approximate time and hardware to generate the data:** 20 minutes per simulation on an A100 GPU in double precision. There is a total of 225 simulations, which is approximately 75 hours.\n\n# What is interesting and challenging about the data:\n\n**What phenomena of physical interest are catpured in the data:** How is energy being transferred between scales? How is vorticity coupled to the orientation field? Where does the transition from isotropic state to nematic state occur with the change in alignment (\\\\(\\zeta\\\\)) or dipole strength (\\\\(\\alpha\\\\))?\n\n\n**How to evaluate a new simulator operating in this space:** Reproducing some summary statistics like power spectra and average scalar order parameters. Additionally, being able to accurately capture the phase transition from isotropic to nematic state.\n\nPlease cite the associated paper if you use this data in your research:\n```\n@article{maddu2024learning,\n  title={Learning fast, accurate, and stable closur","encoding":"utf8","key":"README.md","name":"README.md","size":4337,"truncated":true}],"sourceObjectKey":"data/test/active_matter_L_10.0_zeta_1.0_alpha_-1.0.hdf5","starterCell":"from huggingface_hub import hf_hub_download\npath = hf_hub_download(repo_id=\"polymathic-ai/active_matter\", filename=\"data/test/active_matter_L_10.0_zeta_1.0_alpha_-1.0.hdf5\", repo_type=\"dataset\")\nprint(path)","starterSamplePath":"data/test/active_matter_L_10.0_zeta_1.0_alpha_-1.0.hdf5","totalObjects":88},"summary":"Simulations of a continuum theory for the dynamics of N rod-like active particles immersed in a Stokes fluid, a model of biological active matter such as motor-driven filament…","title":"The Well: Active Matter, Rod-like Active Particles in a Stokes Fluid","updatedAt":"2026-11-14T02:39:22.556Z","visibility":"public","websiteUrl":"https://polymathic-ai.org/the_well/"}}