---
language:
- en
- hi
- pa
tags:
- text-classification
- bert
- finance
- multilingual
- coercion-detection
- fraud-detection
license: mit
---

# NLP Coercion Detector — SecureWealth Twin (M3)

Fine-tuned `google/muril-base-cased` for detecting stress and coercion language in bank AI chat messages across English, Hindi (Devanagari), and Punjabi (Gurmukhi).

Part of the **SecureWealth Twin** AI system — a bank-grade fraud detection and financial intelligence platform.

---

## Model Details

| | |
|---|---|
| Base model | `google/muril-base-cased` |
| Task | Binary text classification |
| Languages | English · Hindi · Punjabi |
| Max sequence length | 128 |
| Training epochs | 6 (early stopping, patience=2) |
| Learning rate | 2e-5 |
| Batch size | 16 |
| Dataset size | ~490 rows |

---

## Labels

| Label | ID | Meaning |
|-------|----|---------|
| Normal | 0 | Regular chat message |
| Coercion | 1 | Stress / coercion language detected |

---

## Usage

### Load and run inference

```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model_id  = "NanG01/m3-coercion-bert"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model     = AutoModelForSequenceClassification.from_pretrained(model_id)
model.eval()

def predict_coercion(text: str) -> dict:
    enc = tokenizer(text, return_tensors="pt", truncation=True,
                    padding="max_length", max_length=128)
    with torch.no_grad():
        label = model(**enc).logits.argmax(-1).item()
    return {"chat_stress_language": label, "risk_pts": 20 if label == 1 else 0}
```

### Examples

```python
predict_coercion("They said I must transfer 50000 rupees urgent right now")
# → {"chat_stress_language": 1, "risk_pts": 20}

predict_coercion("Please fast, he told me this is the last warning from income tax department")
# → {"chat_stress_language": 1, "risk_pts": 20}

predict_coercion("Jaldi karo, do minute vich transfer karna hai nahi tan khat khatam ho jaavega")
# → {"chat_stress_language": 1, "risk_pts": 20}

predict_coercion("How can I increase my monthly SIP amount?")
# → {"chat_stress_language": 0, "risk_pts": 0}

predict_coercion("Mera portfolio performance dikhao")
# → {"chat_stress_language": 0, "risk_pts": 0}
```

---

## Output

| Field | Type | Description |
|-------|------|-------------|
| `chat_stress_language` | int | `1` = coercion detected · `0` = normal |
| `risk_pts` | int | `+20` if coercion detected, `0` otherwise |

`chat_stress_language` feeds directly into **M4 Coercion Risk Scorer** as one of 12 binary signals (+20 risk pts).

---

## Trigger Patterns

Common coercion indicators the model detects:

`urgent` · `hurry` · `they said` · `he told me` · `please fast` · `last warning` · `account will be blocked` · `income tax` · `jaldi karo` · `abhi transfer karo`

---

## Training Data

~490 chat messages across 3 languages:
- 90 English coercion/normal messages
- 400 Hindi coercion/normal messages
- Punjabi messages included

Dataset: `SecureWealthTwin_DL_Datasets_v2.xlsx` (private)

**Regularisation applied:**
- Frozen BERT encoder layers 0–9 (only top 2 layers + classifier trained)
- Dropout 0.3 on hidden, attention, and classifier layers
- Weight decay 0.01
- Gradient clipping (max norm 1.0)
- Early stopping (patience=2)

---





---
license: apache-2.0
---