Add Croissant RAI metadata

README.md

@@ -49,6 +49,10 @@ These datasets are intended for:
 
 They are not intended as general-purpose training corpora.
 
+## Croissant metadata
+
+Hugging Face auto-generates core Croissant metadata for supported dataset repos. Because this artifact is distributed primarily as HDF5 plus CSV/JSON sidecars, the release also includes a self-contained Croissant JSON-LD file with core file metadata and Responsible AI fields at `metadata/croissant_rai.jsonld`.
+
 ## Companion assets
 
 The associated code, manifests, and reviewer notebooks live in the separate GitHub reproducibility repo:

metadata/croissant_rai.jsonld

@@ -0,0 +1,173 @@
+{
+  "@context": {
+    "@language": "en",
+    "@vocab": "https://schema.org/",
+    "citeAs": "cr:citeAs",
+    "column": "cr:column",
+    "conformsTo": "dct:conformsTo",
+    "cr": "http://mlcommons.org/croissant/",
+    "data": {
+      "@id": "cr:data",
+      "@type": "@json"
+    },
+    "dataType": {
+      "@id": "cr:dataType",
+      "@type": "@vocab"
+    },
+    "dct": "http://purl.org/dc/terms/",
+    "equivalentProperty": "cr:equivalentProperty",
+    "examples": {
+      "@id": "cr:examples",
+      "@type": "@json"
+    },
+    "extract": "cr:extract",
+    "field": "cr:field",
+    "fileObject": "cr:fileObject",
+    "fileProperty": "cr:fileProperty",
+    "fileSet": "cr:fileSet",
+    "format": "cr:format",
+    "includes": "cr:includes",
+    "isLiveDataset": "cr:isLiveDataset",
+    "jsonPath": "cr:jsonPath",
+    "key": "cr:key",
+    "md5": "cr:md5",
+    "parentField": "cr:parentField",
+    "path": "cr:path",
+    "rai": "http://mlcommons.org/croissant/RAI/",
+    "recordSet": "cr:recordSet",
+    "references": "cr:references",
+    "regex": "cr:regex",
+    "repeated": "cr:repeated",
+    "replace": "cr:replace",
+    "samplingRate": "cr:samplingRate",
+    "sc": "https://schema.org/",
+    "separator": "cr:separator",
+    "source": "cr:source",
+    "subField": "cr:subField",
+    "transform": "cr:transform"
+  },
+  "@type": "Dataset",
+  "name": "InformedXRD Benchmarks",
+  "alternateName": "informedxrd/informedxrd-benchmarks",
+  "description": "Frozen real PXRD benchmark payloads, reviewer examples, and metadata for the InformedXRD NeurIPS Evaluations & Datasets artifact release.",
+  "url": "https://huggingface.co/datasets/informedxrd/informedxrd-benchmarks",
+  "license": "https://creativecommons.org/licenses/by/4.0/",
+  "version": "0.1.0",
+  "datePublished": "2026-04-30",
+  "creator": {
+    "@type": "Organization",
+    "name": "InformedXRD"
+  },
+  "keywords": [
+    "materials-science",
+    "x-ray-diffraction",
+    "crystallography",
+    "benchmark",
+    "powder-diffraction",
+    "extinction-group-classification"
+  ],
+  "citeAs": "InformedXRD Benchmarks. Hugging Face dataset repository: https://huggingface.co/datasets/informedxrd/informedxrd-benchmarks",
+  "conformsTo": [
+    "http://mlcommons.org/croissant/1.0",
+    "http://mlcommons.org/croissant/RAI/1.0"
+  ],
+  "isLiveDataset": false,
+  "distribution": [
+    {
+      "@type": "cr:FileObject",
+      "@id": "rruff-473-hdf5",
+      "name": "RRUFF_option1_473_with_buckets_maxnorm.hdf5",
+      "description": "Frozen RRUFF-473 benchmark payload with quality buckets and normalized PXRD intensities.",
+      "contentUrl": "benchmark_payloads/RRUFF_option1_473_with_buckets_maxnorm.hdf5",
+      "encodingFormat": "application/x-hdf5",
+      "contentSize": "6.5 MiB",
+      "sha256": "836d30a81879a56afab635cea149d405fb7e2174f972514804f85b470791a5d8"
+    },
+    {
+      "@type": "cr:FileObject",
+      "@id": "rruff-325-hdf5",
+      "name": "RRUFF_usable_plus_recoverable_325_with_labels_maxnorm.hdf5",
+      "description": "Frozen RRUFF-325 usable/recoverable benchmark payload with normalized PXRD intensities and labels.",
+      "contentUrl": "benchmark_payloads/RRUFF_usable_plus_recoverable_325_with_labels_maxnorm.hdf5",
+      "encodingFormat": "application/x-hdf5",
+      "contentSize": "21 MiB",
+      "sha256": "b3dfbc266120668b53bed0db91e9b02a518264ef40061beabd389999c8d9d7f3"
+    },
+    {
+      "@type": "cr:FileObject",
+      "@id": "reviewer-subset-manifest",
+      "name": "manifest.json",
+      "description": "Manifest for the compact reviewer subset.",
+      "contentUrl": "reviewer_subset/manifest.json",
+      "encodingFormat": "application/json",
+      "contentSize": "4 KiB",
+      "sha256": "9a946f3ea290e473462ec5959b58b57964f6b399c97234cde099fe576038ff0a"
+    },
+    {
+      "@type": "cr:FileObject",
+      "@id": "reviewer-case-metadata",
+      "name": "reviewer_case_metadata.csv",
+      "description": "Case metadata for the compact reviewer examples.",
+      "contentUrl": "reviewer_subset/reviewer_case_metadata.csv",
+      "encodingFormat": "text/csv",
+      "contentSize": "4 KiB",
+      "sha256": "97264002a5867a55f884ee5093d3d4d00e624bc12d5977fe735df3410687340d"
+    },
+    {
+      "@type": "cr:FileObject",
+      "@id": "release-manifest",
+      "name": "MANIFEST.csv",
+      "description": "Release-facing artifact manifest.",
+      "contentUrl": "metadata/MANIFEST.csv",
+      "encodingFormat": "text/csv",
+      "contentSize": "8 KiB",
+      "sha256": "82d172ad56ceff8f05a4cdf77ed8ac7682cc6861ba7a2eda6186d76ab7d05031"
+    },
+    {
+      "@type": "cr:FileObject",
+      "@id": "dataset-manifest",
+      "name": "dataset_manifest.csv",
+      "description": "Dataset and external-asset manifest for the companion reproducibility surface.",
+      "contentUrl": "metadata/dataset_manifest.csv",
+      "encodingFormat": "text/csv",
+      "contentSize": "12 KiB",
+      "sha256": "3bd71fa5df24933ae0e00e84761bdf9e133890cfd9cebca928c989b21d67e6d7"
+    }
+  ],
+  "rai:dataCollection": "The dataset packages two frozen real powder X-ray diffraction benchmark payloads derived from curated RRUFF mineral reference patterns. RRUFF-473 retains hard real-data cases for broad robustness analysis, while RRUFF-325 is a cleaner downstream subset for controlled calibration and error-structure analysis. The release also includes a compact reviewer subset, precomputed inference artifacts, and manifests documenting the benchmark definition.",
+  "rai:dataCollectionType": [
+    "Secondary Data analysis",
+    "Manual Human Curator",
+    "Physical data collection"
+  ],
+  "rai:dataCollectionRawData": "The raw experimental PXRD records originate from RRUFF mineral powder diffraction measurements. The released benchmark payloads are derived evaluation artifacts and do not include the broader training corpora, the PRX archival release, or unrelated private intermediate curation outputs.",
+  "rai:dataPreprocessingProtocol": [
+    "Patterns were converted into fixed evaluation HDF5 payloads with normalized intensity arrays and benchmark labels.",
+    "Benchmark membership was frozen using documented curation rules, including quality strata and extinction-group labels.",
+    "RRUFF-325 was derived as the cleaner usable/recoverable subset from the broader RRUFF-473 benchmark surface.",
+    "Reviewer subset CSV and JSON files were exported as compact examples plus precomputed inference summaries."
+  ],
+  "rai:dataAnnotationProtocol": "The benchmark labels are crystallographic extinction-group labels derived from curated crystal-structure metadata and canonical extinction-group mappings. Quality strata and benchmark subset membership are fixed release metadata, not crowd annotations. No demographic or human-subject annotation labels are included.",
+  "rai:machineAnnotationTools": [
+    "Python benchmark-curation and evaluation scripts from the companion public reproducibility repository.",
+    "Crystallographic lookup tables and extinction-group topology assets included in the companion release metadata."
+  ],
+  "rai:dataUseCases": [
+    "Evaluation of real PXRD extinction-group classification.",
+    "Hierarchy-aware rescoring and error analysis for crystallographic symmetry predictions.",
+    "Benchmarking Top-k accuracy, prior baselines, calibration, and stratified evaluation claims."
+  ],
+  "rai:dataBiases": [
+    "The benchmarks inherit coverage biases from RRUFF and from mineral samples with available curated diffraction and structure metadata.",
+    "The class distribution is not intended to represent all inorganic crystals, all minerals, or all experimental PXRD conditions.",
+    "RRUFF-325 is deliberately cleaner than RRUFF-473 and should not be interpreted as a complete hard-case benchmark."
+  ],
+  "rai:dataLimitations": [
+    "The dataset is intended for evaluation, not as a general-purpose training corpus.",
+    "Results on these benchmarks may not generalize to arbitrary instruments, sample preparations, radiation sources, or non-mineral materials.",
+    "The release fixes benchmark metadata for reproducibility; it does not claim invariance to every possible refinement workflow or curation policy."
+  ],
+  "rai:personalSensitiveInformation": "None known. The released artifacts describe mineral diffraction patterns, crystallographic labels, benchmark metadata, and model-evaluation outputs; they do not contain human-subject records or personally identifiable information.",
+  "rai:dataSocialImpact": "The dataset supports reproducible evaluation of ML methods for crystallographic PXRD classification. Potential risks are primarily scientific misuse risks, such as overclaiming model generalization beyond the benchmark scope or treating evaluation payloads as production-ready training data.",
+  "rai:dataReleaseMaintenancePlan": "This NeurIPS artifact release is intended to be frozen for review and reproducibility. Future corrections, if any, should be versioned explicitly and should not silently replace the benchmark payloads used for the submitted paper. The companion code repository and Hugging Face model repository provide the linked reproducibility surface."
+}