PASC: Pipeline-Aware Conformal Prediction with Joint Coverage Guarantees for Multi-Stage NLP and LLM Pipelines

Varun Kotte

arXiv:2605.18812 · 2026-05-20 공개 · arXiv · PDF

conformal-prediction llm-pipelines multi-stage ner ned entity-typing co-nll-2003 wnut-17

Abstract

Modern NLP and LLM systems are pipelines: named entity recognition (NER) -> entity disambiguation (NED) -> entity typing, retrieval-augmented generation (retriever -> reader), and agentic chains of planner -> tool -> critic. Errors compound across stages, but existing uncertainty quantification methods either calibrate each stage independently (no joint coverage) or apply a Bonferroni union bound (joint coverage, but conservative). We present PASC (Pipeline-Aware Split Conformal), which reduces multi-stage joint coverage to a single scalar conformal prediction problem on the joint maximum nonconformity score. PASC provides a finite-sample distribution-free guarantee that all K stages are simultaneously covered with probability at least 1 - alpha, and is nearly tight up to a 1/(n+1) factor. On a three-stage NER -> NED -> entity-typing pipeline over CoNLL-2003, PASC achieves 96.4% end-to-end coverage versus 93.4% for Bonferroni and 86.5% for independent CP, at identical average prediction set size (1.083). Under distribution shift to WNUT-17 Twitter and WikiNEuRal Wikipedia data, PASC empirically maintains the target coverage in the tested shift settings while independent CP collapses to 59%. PASC requires a single quantile computation, runs 1.7x faster than Bonferroni, and scales to K = 6 stages where independent CP drops to 0.53 end-to-end coverage. The same joint-maximum-score reduction applies directly to compound LLM systems and agent pipelines.

한국어 요약

📋 한 줄 요약

**[Conformal Prediction / 파이프라인 보장]** 다단계 NLP·LLM 파이프라인의 결합 커버리지 보장을 단일 스칼라 conformal 문제로 환원하는 PASC 제안.

🎯 핵심 기여도

💡 핵심 아이디어

다단계 파이프라인의 joint coverage는 단계별 보정을 합치는 union bound가 아니라 "각 단계의 최대 nonconformity"라는 하나의 스칼라에 대해서만 보정하면 충분하다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: 다단계 RAG·에이전트 시스템에 적용 가능한 실용적·이론적으로 빡빡한 결합 커버리지 도구 제공. **한계**: nonconformity 점수의 동질성에 의존하는 빡빡함, 단계 간 강한 상관이 있을 때 최적성 분석은 추가 연구 필요.

🚀 실용적 활용