-
Notifications
You must be signed in to change notification settings - Fork 2
References
The Principles Framework is built upon foundational research and methodologies that have proven effective in dynamic task decomposition, agent generation, and multi-agent systems. Below is a list of key references that have influenced the development and validation of Principles.
-
TDAG Framework and ItineraryBench
- Title: TDAG: Dynamic Task Decomposition and Agent Generation for Enhanced Multi-Agent Systems
- Link: https://arxiv.org/abs/2402.10178
-
Dynamic Role Discovery and Assignment
- Title: Dynamic Role Discovery and Assignment in Multi-Agent Systems for Enhanced Task Performance
- Link: https://link.springer.com/content/pdf/10.1007/s40747-023-01071-x.pdf
-
TASKBENCH: A Comprehensive Benchmark for LLM-based Task Automation
- Title: TASKBENCH: Evaluating the Capabilities of Large Language Models in Task Automation
- Link: https://arxiv.org/abs/2311.18760
-
OpenAI Swarm
- Title: OpenAI Swarm: A Multi-Agent System for Complex Problem Solving
- Link: https://github.com/openai/swarm
-
Breaking Down Complexity: A Journey into Multi-Agent Systems and the Future of Collaborative AI
- Title: Breaking Down Complexity: A Journey into Multi-Agent Systems and the Future of Collaborative AI
- Link: https://medium.com/p/77fd7707bdf5
-
OpenAI Documentation: Comprehensive guides and references for integrating OpenAI's models.
-
GitHub Repositories:
-
Related Frameworks and Benchmarks:
- ReAct: Reason+Act framework for agent-based reasoning.
- P&S, P&E, ADAPT: Various baseline methods for comparison in task execution.
We acknowledge the contributions of researchers and developers whose work has significantly influenced the Principles Framework. Their innovative approaches to task decomposition, agent generation, and multi-agent systems have provided a solid foundation for developing effective and adaptable AI solutions.