@inproceedings{silchenko-2026-coverage,
title = "Coverage Is Not Knowledge: Why Closed-Book Knowledge Base Construction Looks Selection Bound",
author = "Silchenko, Maksim",
editor = "Kalo, Jan-Christoph and
Nakashole, Ndapa and
Razniewski, Simon and
Suchanek, Fabian M. and
Vlachos, Andreas and
McCallum, Andrew",
booktitle = "Proceedings of the 11th Workshop on Automated Knowledge Base Construction ({AKBC} 2026)",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.akbc-1.33/",
pages = "447--455",
ISBN = "979-8-89176-464-4",
abstract = "This paper describes the first place system of the AKBC 2026 shared task on predicting complete knowledge base entries from language models, built on a single 31.3B parameter base model in completion mode with no training and no retrieval. It scores 0.7060 on the official test set; the runner up scores 0.6908, and a composite taking the best per relation score of every other ranked team would reach 0.7011.\par The largest single improvement I found is not an algorithm. Asking the model to recall facts about an entity before committing to an answer, rather than asking for the answer directly, moved hasArea from 0.8100 to 0.8700. Three earlier configurations, across two frames and two aggregation methods, had all returned exactly 0.8100, which I had read as evidence that the answers were absent from the candidate pool. That reading was incorrect.\par The main analytical result is a correction to how this field measures headroom, and it corrects my own earlier conclusion as much as anyone else{'}s. Oracle coverage analyses report that hasCapacity is selection bound: the recitation frame puts an acceptable answer in the pool for 84.5 percent of validation subjects while the aggregator realises 36.1 percent. That inference does not survive a null control. A pool of one hundred guesses spanning one decade contains about twenty one mutually exclusive 5 percent tolerance windows, so it contains something acceptably close to almost any plausible target whether or not the model knows the answer. Permuting gold values across subjects measures that coincidence rate directly: 0.355 for hasCapacity against 0.031 for hasArea. Chance corrected, the findable headroom beyond the rank one answer falls from 0.443 to 0.203. Three instruments that are not functions of vote frequency then agree that the relation is knowledge bound."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="silchenko-2026-coverage">
<titleInfo>
<title>Coverage Is Not Knowledge: Why Closed-Book Knowledge Base Construction Looks Selection Bound</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maksim</namePart>
<namePart type="family">Silchenko</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 11th Workshop on Automated Knowledge Base Construction (AKBC 2026)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jan-Christoph</namePart>
<namePart type="family">Kalo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ndapa</namePart>
<namePart type="family">Nakashole</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simon</namePart>
<namePart type="family">Razniewski</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fabian</namePart>
<namePart type="given">M</namePart>
<namePart type="family">Suchanek</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andreas</namePart>
<namePart type="family">Vlachos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andrew</namePart>
<namePart type="family">McCallum</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-464-4</identifier>
</relatedItem>
<abstract>This paper describes the first place system of the AKBC 2026 shared task on predicting complete knowledge base entries from language models, built on a single 31.3B parameter base model in completion mode with no training and no retrieval. It scores 0.7060 on the official test set; the runner up scores 0.6908, and a composite taking the best per relation score of every other ranked team would reach 0.7011.\par The largest single improvement I found is not an algorithm. Asking the model to recall facts about an entity before committing to an answer, rather than asking for the answer directly, moved hasArea from 0.8100 to 0.8700. Three earlier configurations, across two frames and two aggregation methods, had all returned exactly 0.8100, which I had read as evidence that the answers were absent from the candidate pool. That reading was incorrect.\par The main analytical result is a correction to how this field measures headroom, and it corrects my own earlier conclusion as much as anyone else’s. Oracle coverage analyses report that hasCapacity is selection bound: the recitation frame puts an acceptable answer in the pool for 84.5 percent of validation subjects while the aggregator realises 36.1 percent. That inference does not survive a null control. A pool of one hundred guesses spanning one decade contains about twenty one mutually exclusive 5 percent tolerance windows, so it contains something acceptably close to almost any plausible target whether or not the model knows the answer. Permuting gold values across subjects measures that coincidence rate directly: 0.355 for hasCapacity against 0.031 for hasArea. Chance corrected, the findable headroom beyond the rank one answer falls from 0.443 to 0.203. Three instruments that are not functions of vote frequency then agree that the relation is knowledge bound.</abstract>
<identifier type="citekey">silchenko-2026-coverage</identifier>
<location>
<url>https://aclanthology.org/2026.akbc-1.33/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>447</start>
<end>455</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Coverage Is Not Knowledge: Why Closed-Book Knowledge Base Construction Looks Selection Bound
%A Silchenko, Maksim
%Y Kalo, Jan-Christoph
%Y Nakashole, Ndapa
%Y Razniewski, Simon
%Y Suchanek, Fabian M.
%Y Vlachos, Andreas
%Y McCallum, Andrew
%S Proceedings of the 11th Workshop on Automated Knowledge Base Construction (AKBC 2026)
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-464-4
%F silchenko-2026-coverage
%X This paper describes the first place system of the AKBC 2026 shared task on predicting complete knowledge base entries from language models, built on a single 31.3B parameter base model in completion mode with no training and no retrieval. It scores 0.7060 on the official test set; the runner up scores 0.6908, and a composite taking the best per relation score of every other ranked team would reach 0.7011.\par The largest single improvement I found is not an algorithm. Asking the model to recall facts about an entity before committing to an answer, rather than asking for the answer directly, moved hasArea from 0.8100 to 0.8700. Three earlier configurations, across two frames and two aggregation methods, had all returned exactly 0.8100, which I had read as evidence that the answers were absent from the candidate pool. That reading was incorrect.\par The main analytical result is a correction to how this field measures headroom, and it corrects my own earlier conclusion as much as anyone else’s. Oracle coverage analyses report that hasCapacity is selection bound: the recitation frame puts an acceptable answer in the pool for 84.5 percent of validation subjects while the aggregator realises 36.1 percent. That inference does not survive a null control. A pool of one hundred guesses spanning one decade contains about twenty one mutually exclusive 5 percent tolerance windows, so it contains something acceptably close to almost any plausible target whether or not the model knows the answer. Permuting gold values across subjects measures that coincidence rate directly: 0.355 for hasCapacity against 0.031 for hasArea. Chance corrected, the findable headroom beyond the rank one answer falls from 0.443 to 0.203. Three instruments that are not functions of vote frequency then agree that the relation is knowledge bound.
%U https://aclanthology.org/2026.akbc-1.33/
%P 447-455
Markdown (Informal)
[Coverage Is Not Knowledge: Why Closed-Book Knowledge Base Construction Looks Selection Bound](https://aclanthology.org/2026.akbc-1.33/) (Silchenko, AKBC 2026)
ACL