Key principles that define AI agents:

• Autonomy: AI agents can operate independently, making decisions and taking actions without direct human intervention. 
• Reactivity: AI agents can perceive and respond to changes in their environment.
• Proactiveness: AI agents can take initiative and act to achieve their goals without being explicitly instructed. 
• Social Ability: Some AI agents can interact and communicate with other agents, including humans. 
• Rationality: AI agents aim to make decisions that maximize their chances of achieving their goals. 

Essential components of an AI agent system:

Figure 1.2: Conceptual Framework of LLM-based Agent

Figure 1.2 presents a conceptual framework of LLM-based agent, highlighting three components: brain, perception, and action.

• Brain/Agent Core: This is the central component of the agent, responsible for decision-making and planning. In LLM-based agents, the brain is typically an LLM. The brain/agent core integrates information from other components and makes decisions about which actions to take.
• Perception Module: This module allows the agent to perceive its environment, collecting information through sensors or data sources. It can include text, images, sounds, and other sensory inputs.
• Action Module: This module enables the agent to take actions in its environment. These actions can range from physical movements (in the case of robots) to digital interactions, like sending emails or interacting with websites.
• Memory Module: This component allows the agent to store and retrieve information. It can include short-term memory (for recent events) and long-term memory (for storing knowledge and experiences). Examples of memory systems used in AI agents include vector databases like Pinecone and Chroma.
• Tools: Agents can use a variety of tools to enhance their capabilities. Examples include calculators, APIs, search engines, code interpreters, and other specialized software.