Book, AI

Book Review: Multi-Agent Systems with AutoGen by Victor Dibia

My thoughts after reading this book by Manning Publications

--

Image by Author

This book opened a world to me. Before reading this book I was stuck on RAG (Retrieval Augmented Generation), which is fine for fine tuning LLM models, but is not very suitable for more complex systems.

In the book Multi-Agent Systems with AutoGen, Victor Dibia presents AutoGen, a framework for defining agents that collaborate with each other using Artificial Intelligence. The beauty of the book is that it also explains the theoretical concepts behind this framework and not just the code to implement it.

In practice, an agent is a system capable of carrying out a task, such as extracting insight from data. To make sure that everything works, the framework provides an orchestrator that coordinates the various agents. A multiagent system has four fundamental characteristics:

  • Reasoning, that is, synthesizing new information
  • Acting, that is, perform certain actions
  • Adapting, that is, change based on situations
  • Communicating, that is, the various agents can exchange information.

But I don’t want to reveal too much to you: read the book and you will find the details on what a multiagent architecture is and how to implement it.

The book is also enriched with very accurate and detailed explanatory figures, which help the reader to understand what we are talking about.

Let’s look at the book’s index in more detail:

Part 1: Foundations of Multi-Agent Systems

  1. Understanding multi-agent systems
  2. Building your first multi-agent application
  3. UX considerations for a multi-agent system
  4. Interface agents | Agents that solve tasks by interacting with applications

Part 2: Evaluating and Optimizing Multi-Agent Systems

5. Evaluating multi-agent system performance

6. Optimizing multi-agent systems for performance

7. Open challenges in multi-agent application development

8. Ethics and responsible ai in a multi-agent world

Part 3: Real-World Use Cases and Advanced Topics

9. From design to deployment: real-world multi-agent applications

10. Advanced topics in multi-agent application development

Appendices

Appendix A: AutoGen technical documentation

Appendix B: Community resources: forums, articles, and tutorials

Appendix C: Glossary of terms in multi-agent systems

The book in Summary

Below, I summarize some features of the book:

  • Author: Victor Dibia
  • Title: Multi-Agent Systems with AutoGen
  • Publisher: Manning Publications
  • Number of Chapters: 10
  • Who can read this book: developers and programmers wanting to learn MultiAgent

Other books to add to your bookshelf…

Bonus

Recently I have attended the Discovery Science 2024 Conference, where I presented a poster entitled Towards AI-Based Data Analytics for Environmental Monitoring. My co-author and I proposed a system using RAG to extract insights from environmental data automatically. The figure below shows the architecture:

Image by Author

After reading the book Multi-Agent Systems with AutoGen, the next step will be to implement the Testing Phase through AutoGen. In the meantime, you can download the code of this system from here.

Just for curiosity, the poster had an interactive spinning wheel to see different steps of the system. Click here to learn how to implement the wheel in your next poster.

--

--

Angelica Lo Duca
Angelica Lo Duca

Written by Angelica Lo Duca

Researcher | +1M Views | I write on Data Science, Python, Tutorials, and, occasionally, Web Applications | Author of Data Storytelling with Altair and AI