#llm
All the notes tagged “llm”.
Digital Twin and MCP: standards for AI in Smart Cities
How Digital Twin use Model Context Protocol (MCP) as standard to enable real-time city and intelligent services to empower scalable and efficient architectures.
Cognitive Functions as Tools: integrating IA to communicate with other architecture components
The Cognitive Functions as Tools connects AI systems with external software components, significantly expanding their scope and problem-solving capabilities.
Agent handoffs patterns: message flow and delegation of responsibilities
Agent handoffs allow for the transfer of tasks between software agents, ensuring that responsibilities are delegated to the most appropriate or available component. This process enhances system flexibility and scalability by enabling specialized agents to handle requests as needed.
Designing autonomous agents: an architectural approach
Through a structured architectural framework, it explores how to develop intelligent agents capable of making decisions and acting independently in dynamic environments. By understanding the key components and characteristics of agents, developers can design more effective and autonomous software systems.
Patterns to mitigate LLM hallucinations
How to mitigate LLM hallucination issues using effective strategies and software design patterns. Implement these techniques to ensure more accurate and reliable language model outputs.
Retrieval-Augmented Generation (RAG) Systems: custom software components and workflows
How Retrieval-Augmented Generation (RAG) systems utilize various components, powered by the latest Large Language Models, to perform specific tasks in a workflow.Python-based components simplify integration, ensure optimal operation, and enhance language processing capabilities