Modeled on Stanford CS329A (Self-Improving AI Agents)

自我改进的 AI 智能体

Self-Improving AI Agents

Hand-build AI agents that make themselves better, along the skeleton of Stanford CS329A

5 labs15 AI-mentored sessions~6 hoursBilingual · EN / 中
Start learning →

About this course

Why can the same model solve problems it previously couldn't — just by trying multiple
times, adding a "grader", or getting tools and memory? This course follows the skeleton
of Stanford CS329A (Self-Improving AI Agents): test-time compute scaling and verifiers,
self-critique, planning, STaR bootstrapping and RL with verifiable rewards, then
open-ended evolution à la AlphaEvolve, deep-research agents, software-engineering
agents, memory, and agentic evaluation. Every session is hands-on with an AI mentor:
formulas typeset, mechanisms drawn, key loops turned into draggable mini-experiments,
and code that really runs in your own sandbox.

What you'll learn

  • Explain the four levers of self-improvement — sampling, verifiers, tool feedback, training bootstrap — and when each applies
  • Run best-of-N with a verifier and explain why coverage scales log-linearly
  • Implement minimal critique→revise and multi-step planning loops, and tell when they help vs fail
  • Use STaR/RLVR to explain how DeepSeek-R1-style reasoners teach themselves
  • Run a mini evolutionary search (LLM mutation + evaluator selection) and map it to AlphaEvolve/FunSearch
  • Describe how SWE agents, memory augmentation and agentic evals (pass^k, time horizon) work in practice

Syllabus

1Foundations: Sample More, Then Grade3 sessions
  • 1AI That Makes Itself Stronger: A Map25 minStart →
  • 2More Tries: Test-Time Scaling and pass@k25 minStart →
  • 3Who Grades the Answers: Verifiers25 minStart →
2Feedback Loops: Tools, Self-Critique & Planning3 sessions
  • 1Give Your Agent Tools: Think, Act, Observe25 minStart →
  • 2Finding Your Own Flaws: Self-Critique25 minStart →
  • 3Many-Step Problems: Decompose and Plan25 minStart →
3Training Bootstrap: Bake Wins Back into the Model3 sessions
  • 1Teaching Itself: The STaR Bootstrap25 minStart →
  • 2An Auto-Grader as the Teacher: RL with Verifiable Rewards25 minStart →
  • 3From Chatbot to Agent: How Post-Training Evolved25 minStart →
4Open-Ended Evolution: Programs That Grow Themselves3 sessions
  • 1Evolutionary Search: Mutate, Select, and Let Programs Grow25 minStart →
  • 2Agents That Edit Themselves: Darwinian Self-Modification25 minStart →
  • 3Ten Out of a Million: Sampling Funnels and Deep-Research Agents25 minStart →
5Agentic Systems: SWE Agents, Memory & Evals3 sessions
  • 1Hand the AI a Real Bug: SWE Agents25 minStart →
  • 2Giving Agents Memory: From Amnesia to Knowing You Better25 minStart →
  • 3One Great Demo Proves Nothing: Evaluating Agent Reliability25 minStart →

Same series · 世界名校知名实验室系列

Flagship University Lab Series

Modeled on Stanford / MIT / Berkeley syllabi — learn from scratch with a mentor Agent guiding you in real time.