Controlling Authority Retrieval: A Missing Retrieval Objective for Authority-Governed Knowledge
TLDR
This paper introduces Controlling Authority Retrieval (CAR) to find active, non-voided documents in authority-governed knowledge bases, outperforming dense retrieval.
Key contributions
- Formalizes Controlling Authority Retrieval (CAR) for domains where later documents void earlier ones, distinct from semantic search.
- Presents Theorem 4 (CAR-Correctness) and Proposition 2 (Scope Identifiability Upper Bound) for CAR.
- Validates CAR with a two-stage retrieval approach across security, legal, and drug regulation datasets.
- Demonstrates two-stage CAR significantly reduces "not patched" claims in RAG compared to dense retrieval.
Why it matters
This paper addresses a critical, often overlooked problem in knowledge retrieval for authoritative domains like law and security. By formalizing Controlling Authority Retrieval (CAR) and demonstrating its effectiveness, it significantly improves the accuracy of information systems. This prevents outdated or superseded information from being presented, which is crucial for reliable decision-making.
Original Abstract
In any domain where knowledge accumulates under formal authority -- law, drug regulation, software security -- a later document can formally void an earlier one while remaining semantically distant from it. We formalize this as Controlling Authority Retrieval (CAR): recovering the active frontier front(cl(A_k(q))) of the authority closure of the semantic anchor set -- a different mathematical problem from argmax_d s(q,d). The two central results are: Theorem 4 (CAR-Correctness Characterization) gives necessary-and-sufficient conditions on any retrieved set R for TCA(R,q)=1 -- frontier inclusion and no-ignored-superseder -- independent of how R was produced. Proposition 2 (Scope Identifiability Upper Bound) establishes phi(q) as a hard worst-case ceiling: for any scope-indexed algorithm, TCA@k <= phi(q) * R_anchor(q), proved by an adversarial permutation argument. Three independent real-world corpora validate the proved structure: security advisories (Dense TCA@5=0.270, two-stage 0.975), SCOTUS overruling pairs (Dense=0.172, two-stage 0.926), FDA drug records (Dense=0.064, two-stage 0.774). A GPT-4o-mini experiment shows the downstream cost: Dense RAG produces explicit "not patched" claims for 39% of queries where a patch exists; Two-Stage cuts this to 16%. Four benchmark datasets, domain adapters, and a single-command scorer are released at https://github.com/andremir/car-retrieval.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.