Vision Language Model Articles Tracker

This repository tracks and displays the latest research papers and articles related to Vision Language Models (VLMs).

Features

• Automatically fetches articles from various sources (arXiv, Papers With Code, research blogs, etc.)
• Calculates attention scores based on recency, keywords, and source
• Displays articles in a user-friendly dashboard
• Updates weekly with the latest research

Dashboard

The dashboard displays:

• Total number of articles
• Unique keywords
• Last added date
• Articles sorted by attention score
• Keyword trends
• Source distribution

Automatic Updates

The system can be set up to automatically update weekly using a cron job.

Setting up the Weekly Cron Job

1. Make sure the update_articles_weekly.sh script is executable:

   chmod +x update_articles_weekly.sh

2. Edit your crontab:

   crontab -e

3. Add a line to run the script weekly (e.g., every Sunday at 2 AM):

   0 2 * * 0 /full/path/to/update_articles_weekly.sh >> /full/path/to/cron_log.txt 2>&1
   

4. Save and exit the editor.

Manual Updates

You can also update the articles manually by running:

python article_fetcher.py

Attention Score Calculation

Attention scores are calculated based on:

• Recency (newer articles get higher scores)
• Keywords (articles with more relevant keywords get higher scores)
• Source (articles from prestigious sources get higher scores)
• Citation velocity (if available)

Troubleshooting

If you encounter issues with dates or attention scores:

1. Run the update script manually to refresh all data:

   ./update_articles_weekly.sh

2. Check the logs for any errors:

   cat update_log.txt

3. If needed, you can run the individual fix scripts:

   python update_attention_scores.py  # Updates attention scores
   python update_article_files.py     # Updates individual article files

What are Vision Language Models?

Vision Language Models (VLMs) are AI systems that can understand and process both visual and textual information, enabling tasks such as:

• Image captioning
• Visual question answering
• Text-to-image generation
• Multimodal understanding and reasoning

Repository Contents

• article_fetcher.py: Python script that fetches the latest VLM articles from various sources
• articles/: Directory containing saved article metadata in JSON format
• requirements.txt: List of Python dependencies
• .github/workflows/: Automated workflows for periodic article fetching

Latest Articles

Articles are updated automatically on a weekly basis. Check the articles/ directory for the latest content or visit the GitHub Pages site for a more user-friendly view.

Sources

This repository fetches articles from the following sources:

• arXiv (Computer Vision and Machine Learning categories)
• Major AI research labs' blogs (OpenAI, Google AI, Meta AI, etc.)
• Top ML/AI conferences (CVPR, ICCV, NeurIPS, ICLR, etc.)
• Medium articles and popular AI blogs

Setup for Local Use

1. Clone this repository:

git clone https://github.com/wangyunai/vision-language-model-articles.git
cd vision-language-model-articles

2. Install dependencies:

pip install -r requirements.txt

3. Run the article fetcher manually:

python article_fetcher.py

How to Contribute

Contributions are welcome! If you'd like to:

• Add new sources for article fetching
• Improve the filtering algorithm
• Fix bugs or enhance functionality

Please submit a pull request or open an issue.

License

MIT