Roh Donghyun

Home

❯

Posts

❯

Insights

❯

LLM 기반 Agent의 확장 — 연세대 LangAGI Lab 기술 세미나 정리

LLM 기반 Agent의 확장 — 연세대 LangAGI Lab 기술 세미나 정리

Properties1
tagsai-agent

Graph View

  • Agent란 무엇인가
  • Primary Extension: Generalist Web Agent
  • Approach 1 — Test-time scaling: Tree Search Agents
  • Approach 2 — Better learning: Web-Shepherd (Process Reward Modeling)
  • Approach 3 — Context engineering: World Model Training
  • Multi-SW: Web Agent를 넘어 GUI Agent로
  • Secondary Extension: Multi-Agent와 Human–AI Collaboration
  • Research Roadmap: 5단계와 Self-Evolving Agents
  • 정리
  • 참고문헌

Created with Quartz v5.0.0 © 2026

  • GitHub
  • Discord Community