Add paper link, GitHub repository, and sample usage to dataset card

README.md

@@ -1,53 +1,52 @@
 ---
-license: other
-pretty_name: WildfireIA Anonymous Benchmark Release
+license: mit
+size_categories:
+- 10K<n<100K
 task_categories:
 - tabular-classification
 - tabular-regression
 - image-classification
 - other
+pretty_name: WildfireIA
 tags:
 - wildfire
 - benchmark
 - geospatial
 - multimodal
 - croissant
-size_categories:
-- 10K<n<100K
 ---
 
-# WildfireIA Anonymous Benchmark Release
+# WildfireIA: A Nationwide Benchmark for Wildfire Initial Attack Failure Prediction
 
-This anonymous release contains the WildfireIA benchmark data used for review.
-WildfireIA is an event-level benchmark for predicting whether a reported
-Natural wildfire escapes initial attack from public information available at
-fire discovery time.
+WildfireIA is an event-level benchmark for predicting whether a reported Natural wildfire escapes initial attack (IA) using public information available at fire discovery time. It aligns 38,128 naturally caused FPA-FOD wildfire events with FIRMS/VIIRS thermal detections, gridMET weather and fire-danger variables, LANDFIRE vegetation, fuel, and topography, OpenStreetMap access features, and WorldPop population density.
 
-## Contents
-
-- `data/canonical/raw_feature_tables/`: canonical benchmark tables. These are
-  the primary dataset artifact. They contain event-level tables, source-level
-  feature tables, patch-level canonical tables, labels, splits, and manifests.
-- Model-ready caches are not included in this compact release; regenerate
-  them deterministically from the canonical tables with `code/dataloader.py`.
-- `code/`: anonymous copies of the cache generation, training, and summary
-  scripts.
-- `metadata/`: release manifest and cache generation commands.
-- `croissant.json`: Croissant metadata with Responsible AI fields.
+- **Paper:** [A Nationwide Benchmark for Wildfire Initial Attack Failure Prediction with Public Environmental Data](https://huggingface.co/papers/2606.15529)
+- **Repository:** [https://github.com/LabRAI/WildfireIA](https://github.com/LabRAI/WildfireIA)
 
 ## Tasks
 
-Task 1 predicts initial attack failure. The sample unit is one FPA-FOD Natural
-wildfire event. Events with final size at most 10 ha are labeled 0, events with
-final size at least 50 ha are labeled 1, and intermediate-size events are
-excluded from the Task 1 supervised split.
+- **Task 1 (Initial Attack Failure):** Predicts whether a wildfire escapes early control. Events with final size at most 10 ha are labeled 0, while events with final size at least 50 ha are labeled 1.
+- **Task 2 (Containment Duration):** Predicts the remaining time-to-containment as a regression target: `log(1 + containment_hours)`.
 
-Task 2 predicts remaining time-to-containment as a regression target,
-`log(1 + containment_hours)`, using the same discovery-time input contract.
+## Sample Usage
 
-## Rebuilding Model-Ready Caches
+### 1. Download the canonical dataset
 
-The canonical tables can regenerate all official model-ready caches:
+You can download the canonical tables using the `huggingface_hub` library:
+
+```python
+from huggingface_hub import snapshot_download
+
+snapshot_download(
+    repo_id="WildfireIA/Anonymous-WildfireIA",
+    repo_type="dataset",
+    local_dir="hf_data",
+)
+```
+
+### 2. Rebuilding Model-Ready Caches
+
+The canonical tables can regenerate all official model-ready caches (tabular, temporal, spatial, or spatiotemporal) using the provided `dataloader.py` script from the GitHub repository:
 
 ```bash
 python code/dataloader.py \
@@ -61,13 +60,25 @@ python code/dataloader.py \
   --overwrite
 ```
 
-Additional ablation caches are generated by changing `--input_protocol` and
-`--weather_days`; see `metadata/cache_generation_commands.md`.
+## Contents
+
+- `data/canonical/raw_feature_tables/`: Canonical benchmark tables containing event-level tables, source-level feature tables, patch-level canonical tables, labels, and splits.
+- `code/`: Scripts for cache generation, training, and summarization.
+- `metadata/`: Release manifest and cache generation commands.
+- `croissant.json`: Croissant metadata with Responsible AI fields.
 
 ## Responsible Use
 
-The benchmark is intended for reproducible scientific comparison and ablation
-analysis. It should not be used as a standalone operational dispatch system
-without agency validation. The data are public-source derived, but they include
-wildfire locations, fire-station locations, roads, population density, and other
-geospatial context.
+The benchmark is intended for reproducible scientific comparison and ablation analysis. It should not be used as a standalone operational dispatch system without agency validation. The data are public-source derived, but they include wildfire locations, fire-station locations, roads, population density, and other geospatial context.
+
+## Citation
+
+```bibtex
+@misc{xu2026wildfireia,
+  title={A Nationwide Benchmark for Wildfire Initial Attack Failure Prediction with Public Environmental Data},
+  author={Xu, Runyang and Cheng, Xueqi and Dong, Yushun},
+  year={2026},
+  note={Preprint},
+  url={https://github.com/LabRAI/WildfireIA}
+}
+```