Welcome to the world of Insight! This remarkable autonomous AI system is designed to revolutionize medical research by coordinating tasks between various agents, utilizing APIs to access vast information, and allowing you to interact with your data seamlessly. In this blog, we’ll guide you on how to set up and start using Insight for your research tasks.

Understanding Insight’s Architecture

Think of the Insight system as a well-organized factory:

• The Boss Agent is the manager, laying out the objectives and creating a task list based on completed tasks and their results.
• Worker Agents are the diligent employees who pick tasks from the queue and complete them, drawing context from the Llama Index (imagine it as a comprehensive knowledge database).
• Two API workers (PubMed API and MyGene API) act like specialized technicians, fetching valuable information as needed.
• The factory produces outputs, documenting results, summaries, and key findings systematically.

Each component collaborates to ensure that the workflow is smooth, efficient, and enhances the quality of your research outputs.

Getting Started

Follow these steps to kickstart your journey with Insight:

1. Sign up for OpenAI.
2. Expose the following environment variable:
   • OPENAI_API_KEY

   OR

   • Add your API key to the config file. IMPORTANT: If you do this, ensure you do not commit them with any version control system!
3. Run the command: pip install -r requirements.txt
4. Start the application with: python main.py

Output Expectations

Once everything is running, the program will collect results from each completed task and save them in an output directory named out. You’ll also find a markdown file, key_findings.md, that elegantly summarizes your research results using various commands:

• A high-level summary of all the data.
• A list of main points covered by the data.
• Key insights derived from the data.
• Creative hypotheses inspired by the data.
• A discussion of potential research directions.
• A detailed description of key findings.

Troubleshooting

Setbacks can occur during setup or usage of Insight. Here are some troubleshooting ideas:

• If you encounter issues with the API key, double-check that it is correctly added to your environment variables or the config file.
• If the system isn’t responding as expected, ensure you’re in the right directory when running commands.
• Keep an eye on the cost! Insights note that executing tasks for a few minutes should only be a few cents, but be cautious, especially if using powerful models like GPT-4.
• If you find the application isn’t functioning correctly, consider reviewing the requirements.txt for any missing dependencies.

For more insights, updates, or to collaborate on AI development projects, stay connected with fxis.ai.

Conclusion

In summary, Insight is an extraordinary tool for medical research, making complex tasks manageable and efficient through its thoughtfully designed architecture. By leveraging its features, you can extract invaluable insights to push the boundaries of scientific discovery. At fxis.ai, we believe that such advancements are crucial for the future of AI, as they enable more comprehensive and effective solutions. Our team is continually exploring new methodologies to push the envelope in artificial intelligence, ensuring that our clients benefit from the latest technological innovations.