Fake News Detection in Financial News
Develop a generative ai app to detect fake news in financial news articles related to cryptocurrency and Initial Coin Offerings (ICOs). Focus on identifying articles containing misleading information, unsubstantiated claims, or manipulative language. The app should be trained on a diverse dataset of real and fake news articles and should be robust enough to handle various writing styles and deceptive techniques. Evaluate performance using standard metrics like precision, recall, and F1-score.
What you are building
Build a machine learning model to detect fake news in financial news articles about cryptocurrencies and ICOs.
How work is evaluated
The model will be evaluated based on its accuracy, precision, recall, and F1-score on a held-out test set.
Datasets
Public datasets, plus the private datasets you have access to.
Learning goals
Explore techniques to improve model robustness to different writing styles and deceptive techniques.
Train a classification model (e.g., using BERT, RoBERTa) to distinguish between real and fake news articles.
Evaluate the model's performance using appropriate metrics (e.g., precision, recall, F1-score, AUC).
Gather and preprocess a dataset of financial news articles, labeled as real or fake.
Resources
Sample data for 1 tasks
Challenge setup
The model will be evaluated based on its accuracy, precision, recall, and F1-score on a held-out test set.
Challenge input
Text string containing the news article.
Agent execution
No agent tool is listed for this challenge.
Evaluated output
A binary classification label (0 for fake, 1 for real).
- No evaluation checks are configured.
- Accuracy target: 0.85
- Python execution harness
Shows the first task, up to three tools, and up to three evidence items.
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- No evaluation checks are configured.
Metrics
- Accuracy · Target 0.85
- Precision · Target 0.8
- Recall · Target 0.8
Workflow tasks
- Task 1
Input: Text string containing the news article.
Output: A binary classification label (0 for fake, 1 for real).
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.