Upload README.md with huggingface_hub

README.md

@@ -1,157 +1,123 @@
----
-language:
-- en
-- ko
-license: mit
-task_categories:
-- robotics
-- tabular-classification
-tags:
-- robotics
-- failure-analysis
-- pick-and-place
-- nvidia-isaac-sim
-- franka-panda
-- ur5e
-- physical-ai
-- cross-robot
-size_categories:
-- 10K<n<100K
-configs:
-- config_name: default
-  data_files:
-  - split: franka
-    path: data/franka-*
-  - split: franka_boundary
-    path: data/franka_boundary-*
-  - split: ur5e
-    path: data/ur5e-*
-  - split: train
-    path: data/train-*
-dataset_info:
-  features:
-  - name: friction
-    dtype: float64
-  - name: mass
-    dtype: float64
-  - name: com_offset
-    dtype: float64
-  - name: size
-    dtype: float64
-  - name: ik_noise
-    dtype: float64
-  - name: obstacles
-    dtype: int64
-  - name: shape
-    dtype: string
-  - name: placement
-    dtype: string
-  - name: success
-    dtype: bool
-  - name: failure_type
-    dtype: string
-  - name: cycle_time
-    dtype: float64
-  - name: collision
-    dtype: bool
-  - name: drop
-    dtype: bool
-  - name: grasp_miss
-    dtype: bool
-  - name: fail_prob
-    dtype: float64
-  - name: zone
-    dtype: string
-  - name: sample_idx
-    dtype: int64
-  splits:
-  - name: franka
-    num_bytes: 1165541
-    num_examples: 10000
-  - name: franka_boundary
-    num_bytes: 1154763
-    num_examples: 10000
-  - name: ur5e
-    num_bytes: 1021943
-    num_examples: 10000
-  - name: train
-    num_bytes: 3342247
-    num_examples: 30000
-  download_size: 2723675
-  dataset_size: 6684494
----
-
-# RoboGate Failure Dictionary
-
-A structured database of **20,000** robot AI failure experiments from NVIDIA Isaac Sim physical simulations across **2 robot platforms**.
-
-## Quick Stats
-
-| | Franka Panda | UR5e | Combined |
-|---|---|---|---|
-| **Experiments** | 10,000 | 10,000 | **20,000** |
-| **Success Rate** | 33.3% | 74.3% | 53.8% |
-| **Gripper** | Finger (parallel jaw) | SurfaceGripper (suction) | - |
-| **Failure Modes** | grasp_miss, drop, collision | grasp_miss only | 3 types |
-| **#1 Factor** | friction (r=+0.36) | mass (r=-0.35) | - |
-
-## Splits
-
-- `franka`: 10,000 Franka Panda experiments
-- `ur5e`: 10,000 UR5e experiments
-- `all`: 20,000 combined experiments (with `robot` column)
-
-## Usage
-
-```python
-from datasets import load_dataset
-
-ds = load_dataset("liveplex/robogate-failure-dictionary", split="all")
-print(f"Total: {len(ds)}")
-
-# Cross-robot comparison
-franka = load_dataset("liveplex/robogate-failure-dictionary", split="franka")
-ur5e = load_dataset("liveplex/robogate-failure-dictionary", split="ur5e")
-print(f"Franka: {sum(franka['success'])/len(franka):.1%}")
-print(f"UR5e:   {sum(ur5e['success'])/len(ur5e):.1%}")
-```
-
-## Key Findings
-
-- **2.2x success gap**: UR5e (74.3%) vs Franka (33.3%), z=-58.15, p<0.001
-- **Suction never drops**: UR5e drop rate = 0%; all failures are grasp_miss
-- **Friction is universal**: Both robots fail 80%+ when friction < 0.15
-- **Mass kills suction**: UR5e mass [2.4-3.0) -> 85% failure rate
-
-## Schema
-
-| Field | Type | Description |
-|-------|------|-------------|
-| robot | string | "franka_panda" or "ur5e" |
-| friction | float | Surface friction coefficient |
-| mass | float | Object mass in kg |
-| size | float | Object size in meters |
-| com_offset | float | Center of mass offset |
-| ik_noise | float | IK noise level |
-| obstacles | int | Number of obstacles |
-| shape | string | Object shape |
-| placement | string | Object placement |
-| success | bool | Whether the task succeeded |
-| failure_type | string | Classification of failure |
-| cycle_time | float | Time to complete in seconds |
-| collision | bool | Whether collision occurred |
-| drop | bool | Whether object was dropped |
-| grasp_miss | bool | Whether grasp was missed |
-| zone | string | Risk zone (safe/boundary/danger) |
-
-## Citation
-
-```bibtex
-@dataset{robogate_failure_dict_2026,
-  title={RoboGate Failure Dictionary: 20K Physics-Validated Pick & Place Failure Patterns},
-  author={RoboGate Team},
-  year={2026},
-  url={https://huggingface.co/datasets/liveplex/robogate-failure-dictionary},
-  note={Franka Panda + UR5e, NVIDIA Isaac Sim, Latin Hypercube Sampling}
-}
-```
+---
+license: mit
+task_categories:
+  - robotics
+tags:
+  - failure-analysis
+  - pick-and-place
+  - isaac-sim
+  - franka-panda
+  - ur5e
+  - domain-randomization
+  - latin-hypercube-sampling
+  - adaptive-sampling
+  - physical-ai
+size_categories:
+  - 10K<n<100K
+language:
+  - en
+  - ko
+---
+
+# RoboGate Failure Dictionary
+
+> **30,000 Physics-Validated Pick & Place Failure Patterns across Franka Panda & UR5e**
+
+A structured database of robot AI failure patterns collected from **NVIDIA Isaac Sim** physical simulations using **Two-Stage Adaptive Sampling**. Each experiment records the exact conditions under which a robot succeeded or failed at Pick & Place tasks.
+
+## Quick Stats
+
+| | Franka Uniform | Franka Boundary | UR5e | Combined |
+|---|---|---|---|---|
+| **Experiments** | 10,000 | 10,000 | 10,000 | **30,000** |
+| **Success Rate** | 33.3% | 63.8% | 74.3% | — |
+| **Franka Combined** | — | 48.6% | — | — |
+| **Risk Model AUC** | 0.65 | **0.777** | — | 0.777 |
+| **Sampling** | Uniform LHS | Boundary LHS | Uniform LHS | Two-Stage |
+
+## Key Findings
+
+- **friction × mass interaction z = -10.00** — strongest predictor of failure
+- **Friction threshold: 0.492 ± 0.031** — below this, failure cascades
+- **Mass > 0.93 kg** → Both robots fail at **< 40%** SR (universal danger zone)
+- **Boundary equation:** μ*(m) = (1.469 + 0.419m) / (3.691 - 1.400m)
+- **AUC improved 0.65 → 0.777** (+19.5%) with boundary-focused sampling
+- **Failure mode transition:** friction↓ → timeout → collision → grasp_miss
+
+## Two-Stage Adaptive Sampling
+
+**Stage 1 — Uniform Exploration (20,000)**
+- Franka Panda 10K + UR5e 10K
+- Latin Hypercube Sampling for uniform parameter space coverage
+- Identified boundary regions and initial risk model (AUC 0.65)
+
+**Stage 2 — Boundary-Focused (10,000)**
+- Franka Panda only, targeting boundary/transition regions
+- Concentrated sampling near friction threshold 0.492
+- Revealed failure mode transitions invisible to uniform sampling
+- Boosted Risk Model AUC to 0.777 (+19.5%)
+
+## Universal Danger Zones (mass > 0.93 kg)
+
+| Mass Range | Franka SR | UR5e SR |
+|---|---|---|
+| 0.93 – 1.23 kg | 21.4% | 30.9% |
+| 1.23 – 1.52 kg | 14.9% | 25.3% |
+| 1.52 – 1.82 kg | 12.5% | 28.9% |
+| 1.82 – 2.11 kg | 6.6% | 28.1% |
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+ds = load_dataset("liveplex/robogate-failure-dictionary")
+print(ds["train"][0])
+
+# Filter danger zones
+danger = ds["train"].filter(lambda x: x["zone"] == "danger")
+print(f"Danger zones: {len(danger)}")
+```
+
+## Parameter Space
+
+| Parameter | Range | Scale | Paper |
+|-----------|-------|-------|-------|
+| friction | 0.05 – 1.2 | log-uniform | SIMPLER 2024 |
+| mass | 0.05 – 2.0 kg | log-uniform | SIMPLER 2024 |
+| com_offset | 0.0 – 0.40 | uniform | Suction Grasp 2025 |
+| size | 0.02 – 0.12 m | uniform | SIMPLER 2024 |
+| ik_noise | 0.0 – 0.04 rad | uniform | ICRA Sim2Real 2025 |
+| obstacles | 0 – 4 | integer | RoboFAC 2025 |
+| shape | 5 types | categorical | Grasp Anything 2024 |
+| placement | 14 types | categorical | ALEAS 2025 |
+
+## Research Foundations
+
+| Design Choice | Paper | Year |
+|---|---|---|
+| Two-Stage Adaptive Sampling | ALEAS | 2025 |
+| friction × mass interaction | SIMPLER | CoRL 2024 |
+| Failure taxonomy | RoboFAC | NeurIPS 2025 |
+| Cross-robot validation | RoboMIND | RSS 2025 |
+| UR-specific failures | Guardian | ICRA 2025 |
+| Confidence intervals | SureSim | Badithela 2025 |
+| GPU simulation | Isaac Lab | NVIDIA 2025 |
+| Grasp evaluation | Isaac Sim Grasping SDG | NVIDIA 2025 |
+
+## Citation
+
+```bibtex
+@dataset{robogate_failure_dictionary_2026,
+  title={RoboGate Failure Dictionary: 30K Physics-Validated Pick & Place Failure Patterns},
+  author={RoboGate Team},
+  year={2026},
+  url={https://huggingface.co/datasets/liveplex/robogate-failure-dictionary},
+  note={Franka Panda + UR5e, Two-Stage Adaptive Sampling, AUC 0.777}
+}
+```
+
+## Links
+
+- **GitHub**: [liveplex-cpu/robogate-failure-dictionary](https://github.com/liveplex-cpu/robogate-failure-dictionary)
+- **RoboGate Platform**: [liveplex-cpu/robogate](https://github.com/liveplex-cpu/robogate)