Track E: Agents & Autonomous Systems
Building systems that act in the world.
Reading Order
-
ReAct: Synergizing Reasoning and Acting paper
-
Toolformer: Language Models Can Teach Themselves to Use Tools paper
-
Voyager: An Open-Ended Embodied Agent paper
-
Generative Agents: Interactive Simulacra paper
Branches
After generative-agents, explore:
-
Language Agent Tree Search paper
-
Reflexion paper
After lats, explore:
-
MAKER: Solving a Million-Step LLM Task paper
After reflexion, explore:
-
MAKER: Solving a Million-Step LLM Task paper
Key outcome: Design agents that reason, use tools, and learn from feedback.