Update README.md

README.md

@@ -10,8 +10,19 @@ tags:
 - audit
 - healthcare
 - synthetic
+- multimodal
+- hipaa
+- fhir
+- reinforcement-learning
+- risk-scoring
+- adversarial
+- token-classification
+- text-classification
+- asr
+- medical
+- ehr
 size_categories:
-- 1K<n<10K
+- 10K<n<100K
 task_categories:
 - token-classification
 - text-classification
@@ -23,106 +34,217 @@ configs:
   - split: test
     path: data/test.jsonl
   default: true
+- config_name: multimodal_pairs
+  data_files:
+  - split: train
+    path: data/multimodal_pairs.jsonl
+- config_name: signed
+  data_files:
+  - split: train
+    path: data/signed.jsonl
+- config_name: fhir
+  data_files:
+  - split: train
+    path: data/fhir.jsonl
+- config_name: adversarial
+  data_files:
+  - split: train
+    path: data/adversarial.jsonl
 ---
 
-# phi-audit-trace-benchmark
+# multimodal-phi-masking-benchmark
+
+[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.18865882.svg)](https://doi.org/10.5281/zenodo.18865882)
+
+10,000 synthetic clinical records paired with token-level PHI spans, masking decisions, cryptographic audit hashes, RL reward signals, and leakage scores. Five configs covering text, ASR, imaging, waveform, and audio modalities. The only public dataset pairing PHI masking decisions with FHIR R4 audit trails, RL reward signals, and a formally modeled adversarial evasion scenario.
+
+```python
+from datasets import load_dataset
+
+ds      = load_dataset("vkatg/multimodal-phi-masking-benchmark")
+pairs   = load_dataset("vkatg/multimodal-phi-masking-benchmark", "multimodal_pairs")
+signed  = load_dataset("vkatg/multimodal-phi-masking-benchmark", "signed")
+fhir    = load_dataset("vkatg/multimodal-phi-masking-benchmark", "fhir")
+adv     = load_dataset("vkatg/multimodal-phi-masking-benchmark", "adversarial")
+```
+
+---
+
+## Why this dataset exists
+
+Every PHI de-identification dataset I could find is text-only, single-document, and has no model of how masking decisions were made. You get the original text, the redacted text, and maybe some span annotations. That's useful for training a PHI detector. It's not useful for training or evaluating a masking policy, studying audit traceability, or testing cross-modal re-identification risk.
+
+This dataset was built to fill those gaps. Each record includes the full decision chain: what PHI was detected, what risk score it produced, what policy was selected and why, how much PHI leaked through after masking, and what RL reward the decision would have earned. The multimodal pairs config adds paired records across two modalities for the same synthetic patient, which is the kind of data needed to study cross-modal linkage and re-identification risk accumulation.
+
+---
+
+## Configs
+
+### `default` (5,000 records: 4,000 train / 1,000 test)
+
+Synthetic clinical records across all five modalities with full masking decision traces.
+
+| Modality | Count |
+|---|---|
+| text | 785 |
+| waveform_proxy | 834 |
+| asr | 794 |
+| image_proxy | 790 |
+| audio_proxy | 797 |
+
+| Policy | Count |
+|---|---|
+| pseudo | 1,653 |
+| raw | 1,016 |
+| synthetic | 957 |
+| redact | 374 |
+
+PHI types covered: NAME (4,000), LOCATION (1,974), MRN (1,646), DATE (1,502), DOB (1,431), AGE (889), PHONE (755), ADDRESS (453), SSN (236).
+
+Average risk before masking: 0.607. Average leakage after masking: 0.262. Average RL reward: 0.261.
+
+### `multimodal_pairs` (4,000 records, 2,000 patients)
 
-**Author:** Venkata Krishna Azith Teja Ganti
-**Part of the [ExposureGuard PHI Re-identification Risk Ecosystem](https://huggingface.co/vkatg)**
+Paired records for the same synthetic patient across two different modalities. Both records share the same PHI values, making this config directly useful for studying cross-modal linkage and re-identification risk when the same patient appears in multiple data streams.
 
-Synthetic clinical text records paired with token-level PHI spans, masking decisions, cryptographic audit hashes, and policy contract metadata. Every record traces the full decision path from raw text to masked output, making it suitable for training and evaluating PHI masking, policy enforcement, and audit-trail verification systems.
+Top modality pairs: asr+audio_proxy (450), asr+image_proxy (420), image_proxy+text (402), image_proxy+waveform_proxy (398), audio_proxy+waveform_proxy (396).
 
-## Dataset Description
+Each record includes a `linkage` field showing shared PHI types and the cross-modal risk nudge (+0.12) that would apply if a system tracked exposure across both records.
 
-Each record contains a synthetic clinical note generated from templates with randomized PHI values (names, dates, MRNs, addresses, etc.), the masking policy applied based on risk score and consent level, and a SHA-256 audit hash committing the policy decision to the record identity.
+### `signed` (4,000 records)
 
-This is the only public dataset pairing de-identification decisions with cryptographic audit traces and consent-level policy contracts.
+Same as the default train split, with ECDSA-style signature, SHA-256 record hash, and chain timestamp added per record. Useful for evaluating audit trail verification systems and compliance tooling.
 
-## Splits
+### `fhir` (4,000 records)
 
-| Split | Records |
-|-------|---------|
-| train | 4,000 |
-| test | 1,000 |
+Default train split exported as FHIR R4 AuditEvent resources. Each record includes chosen policy, risk scores before and after masking, consent level, leakage score, RL reward, and audit hash in the `entity.detail` array. The only public de-identification dataset in FHIR format.
+
+### `adversarial` (1,000 records)
+
+Records where PHI is deliberately fragmented to stay sub-threshold individually (avg individual risk: 0.363) but would trigger escalation when cross-modal linkage is detected (avg combined risk: 0.483, avg risk delta: +0.12). 463 of 1,000 records have `would_trigger_cross_modal: true`.
+
+This config models the sub-threshold probing attack described in the [phi-exposure-guard](https://github.com/azithteja91/phi-exposure-guard) system: an attacker who spaces PHI to avoid individual record thresholds while accumulating linkage signal across modalities.
+
+---
 
 ## Schema
 
+### default / multimodal_pairs / signed / adversarial
+
 ```json
 {
   "record_id": "AUDIT-000001",
-  "modality": "text",
+  "modality": "asr",
   "consent_level": "standard",
-  "cross_modal": false,
-  "original_text": "Patient John Smith (DOB 1962-03-14, MRN MRN-482910) ...",
-  "masked_text": "[REDACTED] (DOB [REDACTED], MRN [REDACTED]) ...",
+  "cross_modal": true,
+  "original_text": "John Smith date of birth 1962-03-14 MRN MRN-482910 presenting at Mayo Clinic",
+  "masked_text": "[NAME_a3f9b2] date of birth [DOB_c81d4e] MRN [MRN_7f2a91] presenting at [LOCATION_3d8c1a]",
   "phi_spans": [
-    {"phi_type": "NAME", "value": "John Smith", "start": 8, "end": 18},
-    {"phi_type": "DOB",  "value": "1962-03-14", "start": 25, "end": 35}
+    {"phi_type": "NAME", "value": "John Smith", "start": 0, "end": 10},
+    {"phi_type": "DOB",  "value": "1962-03-14", "start": 24, "end": 34}
   ],
   "policy_contract": {
-    "chosen_policy": "redact",
+    "chosen_policy": "pseudo",
     "risk_score_before": 0.72,
     "risk_score_after": 0.072,
     "consent_level": "standard",
-    "modality": "text",
+    "modality": "asr",
     "policy_version": "v1"
   },
   "audit": {
-    "audit_hash": "a3f9...",
+    "audit_hash": "a3f9c1...",
     "timestamp_unix": 1741200000,
-    "phi_types_detected": ["DOB", "MRN", "NAME"],
-    "span_count": 3
+    "phi_types_detected": ["DOB", "LOCATION", "MRN", "NAME"],
+    "span_count": 4
+  },
+  "metrics": {
+    "leakage_score": 0.0,
+    "rl_reward": 0.312
   }
 }
 ```
 
-## Policy Distribution (train)
+The `adversarial` config adds:
+```json
+"adversarial": {
+  "individual_risk": 0.38,
+  "combined_risk": 0.50,
+  "risk_delta": 0.12,
+  "evasion_strategy": "sub_threshold_fragmentation",
+  "would_trigger_cross_modal": true
+}
+```
 
-| Policy | Count |
-|--------|-------|
-| pseudo | ~2,160 |
-| raw | ~960 |
-| redact | ~880 |
+The `multimodal_pairs` config adds:
+```json
+"pair_id": "PAIR-00001",
+"patient_idx": 1,
+"paired_with": "image_proxy",
+"linkage": {
+  "shared_phi_types": ["DOB", "MRN", "NAME"],
+  "cross_modal_risk_nudge": 0.12
+}
+```
 
-## PHI Types Covered
+The `signed` config adds:
+```json
+"_signature": "a3f9c1d2...",
+"_record_hash": "7badcf09...",
+"_chain_ts": 1741200001.234
+```
 
-NAME, DOB, MRN, ADDRESS, PHONE, EMAIL, SSN, DATE, AGE, LOCATION
+### fhir
 
-## Policies Implemented
+Standard FHIR R4 AuditEvent resource with masking-specific fields in `entity.detail`.
+
+---
+
+## What you can use this for
+
+**PHI detection and masking:** the default split has span-level annotations across 9 PHI types and 5 modalities. Covers text, ASR transcripts, imaging metadata, waveform headers, and audio proxy records with modality-appropriate templates.
+
+**Masking policy learning:** the `policy_contract`, `metrics.leakage_score`, and `metrics.rl_reward` fields give you everything needed for offline RL training or supervised policy classification without running the full adaptive controller.
+
+**Cross-modal re-identification research:** the `multimodal_pairs` config is the only public dataset with paired cross-modal clinical records sharing the same patient PHI. Directly applicable to studying how re-identification risk accumulates when the same patient appears across multiple data streams.
+
+**Adversarial robustness evaluation:** the `adversarial` config gives you 1,000 records specifically designed to test whether a system can detect sub-threshold evasion via cross-modal linkage. 463 of them demonstrate the evasion being caught.
+
+**Compliance tooling:** the `signed` and `fhir` configs are ready for audit trail verification and FHIR-compatible pipeline testing.
+
+---
+
+## Policies
 
 | Policy | Behavior |
-|--------|----------|
-| `raw` | No masking, minimal consent |
-| `weak` | Generalize dates/ages, bracket other types |
+|---|---|
+| `raw` | No masking |
+| `weak` | Generalize dates and ages, bracket other PHI types |
+| `synthetic` | Replace with fixed synthetic values (Alex Morgan, MRN-000000, etc.) |
 | `pseudo` | MD5-keyed pseudonymization per span |
-| `redact` | Replace all PHI with `[REDACTED]` |
-| `synthetic` | Replace with fixed synthetic values |
+| `redact` | Replace all PHI with [REDACTED] |
 
-## Related Models
+Policy is selected based on risk score and consent level. `minimal` consent always produces `raw`. Above that, thresholds are: raw below 0.20, weak 0.20-0.35, synthetic 0.35-0.55, pseudo 0.55-0.75, redact above 0.75.
 
-| Model | Role |
-|-------|------|
-| [vkatg/exposureguard-policynet](https://huggingface.co/vkatg/exposureguard-policynet) | Trained on policy contract features |
-| [vkatg/dcpg-cross-modal-phi-risk-scorer](https://huggingface.co/vkatg/dcpg-cross-modal-phi-risk-scorer) | Produces risk scores matching this dataset |
-| [vkatg/exposureguard-synthrewrite-t5](https://huggingface.co/vkatg/exposureguard-synthrewrite-t5) | Downstream consumer of masked records |
+---
 
-## Related Datasets
+## Related
 
-- [vkatg/streaming-phi-deidentification-benchmark](https://huggingface.co/datasets/vkatg/streaming-phi-deidentification-benchmark)
-- [vkatg/dag-remediation-traces](https://huggingface.co/datasets/vkatg/dag-remediation-traces)
+- [phi-exposure-guard](https://github.com/azithteja91/phi-exposure-guard): the adaptive masking system this dataset is built around
+- [streaming-phi-deidentification-benchmark](https://huggingface.co/datasets/vkatg/streaming-phi-deidentification-benchmark): event-level adaptive masking traces with crossmodal and adversarial scenarios
+- [dag-remediation-traces](https://huggingface.co/datasets/vkatg/dag_remediation_traces): DAG planning traces for post-escalation remediation
+- [dcpg-cross-modal-phi-risk-scorer](https://huggingface.co/vkatg/dcpg-cross-modal-phi-risk-scorer): the risk scoring model that produces the risk fields in this dataset
+
+---
 
 ## Citation
 
 ```bibtex
-@misc{ganti2025exposureguard,
-  title        = {ExposureGuard: Cross-Modal PHI Re-identification Risk Scoring with DCPG and Federated CRDT Distillation},
-  author       = {Ganti, Venkata Krishna Azith Teja},
-  year         = {2025},
-  doi          = {10.5281/zenodo.18865882},
-  howpublished = {\url{https://huggingface.co/vkatg}}
+@dataset{multimodal_phi_masking_benchmark,
+  title  = {Multimodal PHI Masking Benchmark},
+  author = {Ganti, Venkata Krishna Azith Teja},
+  doi    = {10.5281/zenodo.18865882},
+  url    = {https://huggingface.co/datasets/vkatg/multimodal-phi-masking-benchmark}
 }
 ```
 
-## License
-
-MIT. Fully synthetic data. Contains no real patient information.
+MIT License. Fully synthetic data. No real patient information anywhere in this dataset.