Inspiration
approximately a total number of 2.6 million submissions are available on ArXiv since August 1991, it is even increasing at tempo of 20000 per month. The agent will regularly provide concise and accessible reports to keep the team updated on selected topics. This initiative aims to enhance research efficiency and ensure timely access to cutting-edge knowledge.
What it does
it consists of following workflows: defining research topic, scrapinig papers from arxiv site, tagging arxiv paper with trained model and generating report. The output is a report web app that presents summary of the papers.
How we built it
it build automatially with github actions and deployed on azure web app
Challenges we ran into
customizing language model for tagging arxiv papers, and integration with chatgpt
Accomplishments that we're proud of
it is end to end workflow with minimal human intervention and cloud cost.
What we learned
github copilot is a great tool for synthetic dataset generation and code generation.
What's next for Report Agent Web App
add more features such as PowerPoint generation etc.
Log in or sign up for Devpost to join the conversation.