README.md

---
language:
- en
license: other
library_name: transformers
tags:
- dark-pattern
- dark-pattern-classification
- BERT
- dark-pattern-detection
metrics:
- accuracy
pipeline_tag: text-classification
---

# Model Card for Model ID

<!-- Provide a quick summary of what the model is/does. -->

This modelcard aims to be a base template for new models. It has been generated using [this raw template](https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/modelcard_template.md?plain=1).

## Model Details

### Model Description

<!-- Provide a longer summary of what this model is. -->

- **Developed by:** [Adarsh Maurya]
- **Model type:** [Safetensors-F32]
- **License:** [Other]
- **Finetuned from model:** [google-bert/bert-base-uncased]

### Model Sources [optional]

<!-- Provide the basic links for the model. -->

- **Repository:** [https://github.com/4darsh-Dev/CogniGaurd]
- **Paper [optional]:** [More Information Needed]
- **Demo:** [https://huggingface.co/spaces/4darsh-Dev/dark_pattern_detector_app]

## Uses

<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
1. For Detection of Text Based Dark Patterns.
2. It has been to classify dark patterns in 7 Categories( Urgency, Scarcity, Misdirection, Social-Proof, Obstruction, Sneaking, Forced Action) + Not Dark Pattern.
### Direct Use

<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->

### Usage
This model can be loaded and used with the Transformers library:

```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer

model_name = "your-username/your-model-name"
model = AutoModelForSequenceClassification.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)

# Example usage
text = "Only 2 items left in stock!"
inputs = tokenizer(text, return_tensors="pt")
outputs = model(**inputs)
predictions = outputs.logits.argmax(-1)

```


## How to Get Started with the Model

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

class DarkPatternDetector:
    def __init__(self, model_name):
        self.label_dict = {
            0: "Urgency", 1: "Not Dark Pattern", 2: "Scarcity", 3: "Misdirection",
            4: "Social Proof", 5: "Obstruction", 6: "Sneaking", 7: "Forced Action"
        }
        self.device = 'cuda' if torch.cuda.is_available() else 'cpu'
        print(f"Using device: {self.device}")

        self.model = AutoModelForSequenceClassification.from_pretrained(model_name).to(self.device)
        self.tokenizer = AutoTokenizer.from_pretrained(model_name)

    def predict(self, text):
        inputs = self.tokenizer(text, return_tensors="pt", truncation=True, padding=True).to(self.device)
        
        with torch.no_grad():
            outputs = self.model(**inputs)
            probabilities = torch.nn.functional.softmax(outputs.logits, dim=1)
            predicted_label = torch.argmax(probabilities, dim=1).item()

        return self.label_dict[predicted_label]

# Usage
if __name__ == "__main__":
    # Replace with your Hugging Face model name
    model_name = "your-username/your-model-name"
    detector = DarkPatternDetector(model_name)

    # Example usage
    texts_to_predict = [
        "Only 2 items left in stock!",
        "This offer ends in 10 minutes!",
        "Join now and get 50% off!",
        "By clicking 'Accept', you agree to our terms and conditions."
    ]

    for text in texts_to_predict:
        result = detector.predict(text)
        print(f"Text: '{text}'\nPredicted Dark Pattern: {result}\n")



```

## Training Details

### Training Data

<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

[More Information Needed]

### Training Process
- The model was fine-tuned for 5 epochs on a dataset of 5,000 examples.
- We used the AdamW optimizer with a learning rate of 2e-5.
- The maximum sequence length was set to 256 tokens.
- Training was performed using mixed precision (FP16) for efficiency.


<!-- #### Training Hyperparameters

- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> -->

<!-- #### Speeds, Sizes, Times [optional] -->

<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->

[More Information Needed]

<!-- ## Evaluation -->

<!-- This section describes the evaluation protocols and provides the results. -->


### Testing Data, Factors & Metrics

#### Testing Data

<!-- This should link to a Dataset Card if possible. -->

[More Information Needed]


#### Metrics

Our model's performance is evaluated using the following metrics:

- **Accuracy**: The proportion of correct predictions among the total number of cases examined.
- **Precision**: The ratio of correctly predicted positive observations to the total predicted positive observations.
- **Recall**: The ratio of correctly predicted positive observations to all observations in the actual class.
- **F1-Score**: The harmonic mean of Precision and Recall, providing a single score that balances both metrics.

These metrics were chosen to provide a comprehensive view of the model's performance across different aspects of classification accuracy.

### Results

| Metric     | Score    |
|------------|----------|
| Accuracy   | 0.811881 |
| Precision  | 0.808871 |
| Recall     | 0.811881 |
| F1-Score   | 0.796837 |

Our model demonstrates strong performance across all metrics:

- An accuracy of 81.19% indicates that the model correctly classifies a high proportion of samples.
- The precision of 80.89% shows that when the model predicts a specific dark pattern, it is correct about 81% of the time.
- The recall of 81.19% indicates that the model successfully identifies about 81% of the actual dark patterns in the dataset.
- An F1-Score of 79.68% represents a good balance between precision and recall.


### Summary

These results suggest that the model is effective at detecting and classifying dark patterns, with a good balance between identifying true positives and avoiding false positives.


### Model Architecture and Objective


### Compute Infrastructure

#### Hardware
- GPU: NVIDIA Tesla P100 (16GB VRAM)
- Platform: Kaggle Notebooks

#### Software
- Python 3.10
- PyTorch 1.13.1
- Transformers library 4.29.2
- CUDA 11.6


<!-- ## Citation [optional] -->

<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->

<!-- **BibTeX:**

[More Information Needed]

**APA:**

[More Information Needed] -->

## Model Card Authors

This model card was authored by:

- Adarsh Maurya (CS Student, Keshav Mahavidyala[UOD])


## Model Card Contact

For questions, comments, or feedback about this model, please contact:

- Email: adarsh@onionreads.com
- GitHub: [https://github.com/4darsh-Dev/CogniGaurd](https://github.com/4darsh-Dev/CogniGaurd)
- Twitter: [@4darsh_Dev](https://twitter.com/XYZDarkPatternLab)

For urgent inquiries, don't hesitate to get in touch with the lead researcher:
Mr. Adarsh Maurya
Email: adarsh230427@keshav.du.ac.in