From f350af49a5294f5ef3d5ec35b9376aab2bf796ad Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Roberto=20Carratal=C3=A1?= Date: Wed, 20 Nov 2024 15:42:17 +0100 Subject: [PATCH] agentic integration upgrade --- .../modules/ROOT/pages/05-01-agentic-integrations.adoc | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/content/modules/ROOT/pages/05-01-agentic-integrations.adoc b/content/modules/ROOT/pages/05-01-agentic-integrations.adoc index 9c8e44d..9299966 100644 --- a/content/modules/ROOT/pages/05-01-agentic-integrations.adoc +++ b/content/modules/ROOT/pages/05-01-agentic-integrations.adoc @@ -30,10 +30,10 @@ image::05/05-01-tool-calling.png[Tool Calling] ## Key Features for Custom Agent Architectures -- **Human-in-the-Loop**: Involves human oversight for critical decisions, such as approving actions or editing the agent’s state. This is crucial for sensitive tasks where full automation may not be desirable. -- **Parallelization**: Supports concurrent processing of multiple tasks, optimizing efficiency in complex multi-agent systems. -- **Subgraphs**: Enable modular design by managing isolated state for individual agents, allowing for hierarchical organization and controlled communication. -- **Reflection**: Provides feedback for self-correction, evaluating task completion and improving agent reliability through iterative learning. + - **Human-in-the-Loop**: Involves human oversight for critical decisions, such as approving actions or editing the agent’s state. This is crucial for sensitive tasks where full automation may not be desirable. + - **Parallelization**: Supports concurrent processing of multiple tasks, optimizing efficiency in complex multi-agent systems. + - **Subgraphs**: Enable modular design by managing isolated state for individual agents, allowing for hierarchical organization and controlled communication. + - **Reflection**: Provides feedback for self-correction, evaluating task completion and improving agent reliability through iterative learning. ## Memory and Planning