Leveraging Autonomous Agents on the Web
summary
Demonstrates building AMEE, a goal-based agent that navigates random web-hosted mazes expressed as XML/JSON/RDF over HTTP using ALPS profiles for vocabulary and affordances. Implementation uses a simple wall-following rule rather than ML, runs a 100×100 maze in 94 moves and completes typical mazes in under a second, and publishes code/data in the Amy-Pro GitHub repos. Shows which percepts, actions and state machine rules are required.
AI-written summary · about
chapters (10)
- 0:00Introduction and context
- 7:56Better data beats more data
- 11:08What is an autonomous agent
- 24:51Designing the maze agent
- 36:04Writing the agent code
- 40:20Demo of the maze agent
- 46:01Applying maze logic to the web
- 48:50Why multiple small agents work best
- 50:17Resources and next steps
- 52:04Q&A and closing
related talks
49mAI Agents: Collaboration in Real-Life Applications
32mBuilding Multi-Agent Systems: From Concepts to Collaboration
31mAgentic AI architecture patterns
58mAdvanced Autonomous AI Agents: Memory, Collaboration and Evolution in Action
34mAI Agents: Transforming Autonomous Decision-Making
49mAgile for Agents Managing Robots The Way We Manage Humans
40mAI Agents in Practice
24mAgentic Al in SW Development: Evolving Patterns & Protocols