탐색만 늘린다고 AI가 똑똑해지는 것은 아니다 — Stanford CS329A Part 7
Stanford CS329A Part 7을 따라 AlphaCode, AlphaCode 2, Search-o1, Search-R1을 연결한다. Search의 핵심은 더 많이 찾는 것이 아니라 더 잘 생성·선별·압축하는 데 있다.
Insight Publishing
Search Is Not Enough: How AI Agents Learn to Find the Right Answer — Stanford CS329A Part 7
Stanford CS329A Part 7 connects AlphaCode, AlphaCode 2, Search-o1, and Search-R1 to show why search only helps when generation, selection, and context refinement improve together.
AI는 자기 추론에서 어떻게 다시 학습하는가 — Stanford CS329A Part 6 Train-Time Scaling
Stanford CS329A Part 6를 따라 STaR, DeepSeekMath/GRPO, DAPO를 연결한다. 검증된 reasoning이 어떻게 지속적인 train-time improvement로 바뀌는지 살펴본다.
Train-Time Scaling: How AI Learns from Its Own Reasoning — Stanford CS329A Part 6
Stanford CS329A Part 6 connects STaR, DeepSeekMath/GRPO, and DAPO to show how verified reasoning can be turned into persistent train-time improvement.
AI 에이전트의 능력과 신뢰성은 다르다 — Stanford CS329A Part 5 Agent Evaluation
Stanford CS329A Part 5의 METR, GDPval, DeepScholarBench를 통해 agent capability와 reliability, context, 전문가 수준 산출물 품질이 왜 별개인지 살펴본다.
Capability Is Not Reliability — Stanford CS329A Part 5 Agent Evaluation
Stanford CS329A Part 5 compares METR, GDPval, and DeepScholarBench to show why agent capability, reliability, context, and real-world deliverable quality must be evaluated separately.
AI 에이전트의 피드백은 어디서 와야 하는가 — Stanford CS329A Part 4
Stanford CS329A Part 4를 따라 ReAct, 코드 실행 피드백, Constitutional AI를 하나의 질문으로 연결한다. Self-improving agent는 어떤 교정 신호를 믿어야 하는가?
Where Should an AI Agent Get Its Feedback? — Stanford CS329A Part 4
Stanford CS329A Part 4 connects ReAct, execution feedback, and Constitutional AI to one question: where should a self-improving agent get corrective signals it can actually trust?
Verifier도 틀릴 수 있다 — Stanford CS329A Part 3 Robust Verification
Stanford CS329A Part 3를 따라 outcome verifier, process reward model, Math-Shepherd, weak-verifier ensemble과 distillation까지 검증 기술의 진화를 살펴본다.
When the Verifier Can Be Wrong — Stanford CS329A Part 3
Stanford CS329A Part 3 traces robust verification from outcome verifiers and process reward models to Math-Shepherd, weak-verifier ensembles, and verifier distillation.