Agentic Systems
Created:
(Note: generated by Google Gemini)
Here’s a breakdown of agentic systems, their key features, and their implications:
What are Agentic Systems?
Agentic systems are a category of advanced AI systems focused on autonomy and goal-directed behavior. They possess the following characteristics:
- Goal-Oriented: Agentic systems are designed to understand and pursue complex goals and workflows, often set by humans.
- Autonomous Decision-Making: They can independently decide on the steps or actions needed to achieve the specified goal.
- Planning: Agentic systems break down large goals into smaller, manageable sub-tasks.
- Adaptability: They can react to changing circumstances and adjust their plans and actions to maintain progress towards their goal.
- Contextual Understanding: Many agentic systems are built to process and interpret complex information, including natural language, to make informed decisions.
Examples of Agentic Systems
While still an evolving area of AI, here are some types of systems that exhibit agentic qualities:
- Virtual Assistants and Chatbots: Advanced versions can autonomously handle tasks like scheduling meetings, booking travel, or providing summaries of information from multiple sources.
- Resource Management Systems: These systems can optimize inventory levels, supply chain routes, or manage energy distribution based on changing demands and constraints.
- Algorithmic Trading Systems: These systems can independently analyze market data and execute trades, adapting to market fluctuations while working towards a financial goal.
- Self-Driving Car Technology: These complex systems incorporate numerous agentic components for pathfinding, perception, and decision-making to navigate roads safely.
Implications and Considerations of Agentic Systems
- Increased Efficiency and Productivity: Agentic systems promise to automate complex tasks, freeing human time and resources.
- New Forms of Work Collaboration: Agentic AI may become collaborative partners, capable of understanding human intent and adjusting behavior accordingly.
- Governance and Safety Challenges: Ensuring agentic systems operate within ethical boundaries and don’t lead to unintended consequences is a key area of research and development.
Tao, Wei, Yucheng Zhou, Wenqiang Zhang, and Yu Cheng. “MAGIS: LLM-Based multi-agent framework for GitHub issue resolution,” 2024. https://arxiv.org/abs/2403.17927.
MAGIS: a novel LLM-based Multi-Agent framework for GitHub Issue reSolution, MAGIS, consisting of four kinds of agents customized for the software evolution: Manager, Repository Custodian, Developer, and Quality Assurance Engineer agents. This framework leverages the collaboration of various agents in the planning and coding process to unlock the potential of LLMs to resolve GitHub issues. In experiments, we employ the SWE-bench benchmark to compare MAGIS with popular LLMs, including GPT-3.5, GPT-4, and Claude-2.
Qian, Chen, Xin Cong, Wei Liu, Cheng Yang, Weize Chen, Yusheng Su, Yufan Dang, et al. “Communicative agents for software development,” 2023. https://arxiv.org/abs/2307.07924.
Note: This paper was mentioned in Crewai’s multi ai agent system course on deeplearning.ai
In this paper, we present an innovative paradigm that leverages large language models (LLMs) throughout the entire software development process, streamlining and unifying key processes through natural language communication, thereby eliminating the need for specialized models at each phase. At the core of this paradigm lies ChatDev, a virtual chat-powered software development company that mirrors the established waterfall model, meticulously dividing the development process into four distinct chronological stages: designing, coding, testing, and documenting. Each stage engages a team of “software agents”, such as programmers, code reviewers, and test engineers, fostering collaborative dialogue and facilitating a seamless workflow. The chat chain acts as a facilitator, breaking down each stage into atomic subtasks. This enables dual roles, allowing for proposing and validating solutions through context-aware communication, leading to efficient resolution of specific subtasks. The instrumental analysis of ChatDev highlights its remarkable efficacy in software generation, enabling the completion of the entire software development process in under seven minutes at a cost of less than one dollar. It not only identifies and alleviates potential vulnerabilities but also rectifies potential hallucinations while maintaining commendable efficiency and cost-effectiveness.
Microsoft Autogen – Wu, Qingyun, Gagan Bansal, Jieyu Zhang, Yiran Wu, Shaokun Zhang, Erkang (Eric) Zhu, Beibin Li, Li Jiang, Xiaoyun Zhang, and Chi Wang. AutoGen: Enabling next-gen LLM applications via multi-agent conversation, 2023. https://www.microsoft.com/en-us/research/publication/autogen-enabling-next-gen-llm-applications-via-multi-agent-conversation-framework/.