Update README.md

README.md

@@ -1,90 +1,73 @@
----
-license: cc-by-4.0
-task_categories:
-  - visual-question-answering
-  - question-answering
-language:
-  - en
-tags:
-  - embodied-ai
-  - spatial-reasoning
-  - benchmark
-  - evaluation
-  - indoor-scenes
-  - simulation
-size_categories:
-  - 1K<n<10K
-pretty_name: ESI-Bench
----
-
-# ESI-Bench: Embodied Spatial Intelligence Benchmark
-
-## Dataset Description
-
-ESI-Bench is a comprehensive benchmark for evaluating **Embodied Spatial Intelligence** in AI systems.
-It covers 10 task categories spanning diverse aspects of spatial reasoning, perception, and physical understanding
-in realistic indoor simulation environments.
-
-## Task Categories
-
-| # | Category | Description |
-|---|----------|-------------|
-| 1 | **Action Sequencing** | Inferring the correct order of actions in a scene |
-| 2 | **Cognitive Mapping** | Building and reasoning over mental maps of environments |
-| 3 | **Enumerative Perception** | Counting and enumerating objects or events |
-| 4 | **Metric Comparison** | Comparing distances, sizes, and spatial metrics |
-| 5 | **Perceptual Grounding** | Grounding language to visual perceptual features |
-| 6 | **Physical Dynamics** | Understanding physical cause-and-effect relationships |
-| 7 | **Physical Structure** | Reasoning about physical properties and structures |
-| 8 | **Spatial Relations** | Understanding spatial relationships between objects |
-| 9 | **Specular Reflection** | Reasoning about mirror and reflective surfaces |
-| 10 | **Temporal Understanding** | Understanding temporal order and scene changes |
-
-## Dataset Structure
-
-```
-data/
-├── {task_category}.json          # Top-level index for each task category
-└── {task_category}/
-    └── {subtask_name}/
-        └── {scene_name}/
-            └── *.json            # Individual test cases
-```
-
-## Data Fields
-
-Each JSON file contains structured evaluation data including:
-- **question**: The evaluation question or prompt
-- **answer**: Ground truth answer
-- **metadata**: Scene, subtask, and category information
-
-## Usage
-
-```python
-from datasets import load_dataset
-
-# Load full dataset
-ds = load_dataset("RegulusYin/esi-bench")
-
-# Or load a specific task category index
-import json, requests
-url = "https://huggingface.co/datasets/RegulusYin/esi-bench/resolve/main/data/Action Sequencing.json"
-data = requests.get(url).json()
-```
-
-## License
-
-This dataset is released under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/).
-
-## Citation
-
-If you use ESI-Bench in your research, please cite:
-
-```bibtex
-@dataset{esibench2026,
-  title   = {ESI-Bench: Embodied Spatial Intelligence Benchmark},
-  author  = {RegulusYin et al.},
-  year    = {2026},
-  url     = {https://huggingface.co/datasets/RegulusYin/esi-bench}
-}
-```
+---
+license: cc-by-4.0
+task_categories:
+  - visual-question-answering
+  - question-answering
+language:
+  - en
+tags:
+  - embodied-ai
+  - spatial-reasoning
+  - benchmark
+  - evaluation
+  - indoor-scenes
+  - simulation
+size_categories:
+  - 1K<n<10K
+pretty_name: ESI-Bench
+---
+
+# ESI-Bench: Embodied Spatial Intelligence Benchmark
+
+## Dataset Description
+
+ESI-Bench is a comprehensive benchmark for evaluating **Embodied Spatial Intelligence** in AI systems.
+It covers 10 task categories spanning diverse aspects of spatial reasoning, perception, and physical understanding
+in realistic indoor simulation environments.
+
+## Task Categories
+
+| # | Category | Description |
+|---|----------|-------------|
+| 1 | **Action Sequencing** | Inferring the correct order of actions in a scene |
+| 2 | **Cognitive Mapping** | Building and reasoning over mental maps of environments |
+| 3 | **Enumerative Perception** | Counting and enumerating objects or events |
+| 4 | **Metric Comparison** | Comparing distances, sizes, and spatial metrics |
+| 5 | **Perceptual Grounding** | Grounding language to visual perceptual features |
+| 6 | **Physical Dynamics** | Understanding physical cause-and-effect relationships |
+| 7 | **Physical Structure** | Reasoning about physical properties and structures |
+| 8 | **Spatial Relations** | Understanding spatial relationships between objects |
+| 9 | **Specular Reflection** | Reasoning about mirror and reflective surfaces |
+| 10 | **Temporal Understanding** | Understanding temporal order and scene changes |
+
+## Dataset Structure
+
+```
+data/
+├── {task_category}.json          # Top-level index for each task category
+└── {task_category}/
+    └── {subtask_name}/
+        └── {scene_name}/
+            └── *.json            # Individual test cases
+```
+
+## Data Fields
+
+Each JSON file contains structured evaluation data including:
+- **question**: The evaluation question or prompt
+- **answer**: Ground truth answer
+- **metadata**: Scene, subtask, and category information
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+# Load full dataset
+ds = load_dataset("RegulusYin/esi-bench")
+
+# Or load a specific task category index
+import json, requests
+url = "https://huggingface.co/datasets/RegulusYin/esi-bench/resolve/main/data/Action Sequencing.json"
+data = requests.get(url).json()
+```