From the course: Agentic AI Solution Design Patterns
Join today to access over 25,300 courses taught by industry experts.
Human feedback loop integration
From the course: Agentic AI Solution Design Patterns
Human feedback loop integration
- In the previous video, we explored how an agent can get help from a human when faced with an unforeseen problem. However, even when an agent successfully completes a task or learns a new skill, there's a broader ongoing need for the agent to improve over time. The challenge here is that LLMs often lack subtle visual interpretation or an understanding of human preferences. Without a way to capture and incorporate human judgment, the agent might not be able to truly evolve its effectiveness beyond its programming and training. Now, let's have a look at an example involving a new agentic robot that is working at an assembly line responsible for inspecting toys for defects. Although the robot was trained for this task, it still misses some types of subtle defects, which, therefore, makes it approve toys that can have flaws. For this type of continuous refinement in more nuanced detection tasks, the LLM on its own would not be able to improve its performance over time because its…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.