@inproceedings{asare-aryal-2026-ai4pc,
title = "{AI}4{PC}-{H}oward {U}niversity at {S}em{E}val-2026 Task 5: Calibrated Hybrid Ensembling and Retrieval-Augmented {LLM} Reasoning for Narrative Word-Sense Plausibility",
author = "Asare, Kwaku and
Aryal, Saurav",
editor = "Kochmar, Ekaterina and
Ghosh, Debanjan and
North, Kai and
Komachi, Mamoru",
booktitle = "Proceedings of the 20th {I}nternational {W}orkshop on {S}emantic {E}valuation (2026)",
month = jul,
year = "2026",
address = "San Diego, California, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.semeval-1.326/",
pages = "2592--2596",
ISBN = "979-8-89176-414-9",
abstract = "We present two complementary approaches for rating word-sense plausibility in SemEval-2026 Task 5 (literary homonyms in five-sentence stories). Approach 1 is a retrieve-then-generate pipeline using an open-weight Llama 3.1 70B Instruct model with structured reasoning and a self-correction pass. Approach 2 is a hybrid ensemble that combines API-based LLM prompting with transformer representations and a learned calibration layer trained on the development set. On the development set, Approach 2 achieves Spearman {\ensuremath{\rho}} = 0.7393 (p 10-102) with accuracy 0.8010 (471/588). Approach 1 achieves {\ensuremath{\rho}} = 0.5187 (p 10-65) with accuracy 0.6032 (561/930). We emphasize that Approach 1 does not exceed RoBERTabase in accuracy (0.6032 vs. 0.6410), but provides stronger rank correlation."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="asare-aryal-2026-ai4pc">
<titleInfo>
<title>AI4PC-Howard University at SemEval-2026 Task 5: Calibrated Hybrid Ensembling and Retrieval-Augmented LLM Reasoning for Narrative Word-Sense Plausibility</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kwaku</namePart>
<namePart type="family">Asare</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Saurav</namePart>
<namePart type="family">Aryal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 20th International Workshop on Semantic Evaluation (2026)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Kochmar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Debanjan</namePart>
<namePart type="family">Ghosh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kai</namePart>
<namePart type="family">North</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mamoru</namePart>
<namePart type="family">Komachi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-414-9</identifier>
</relatedItem>
<abstract>We present two complementary approaches for rating word-sense plausibility in SemEval-2026 Task 5 (literary homonyms in five-sentence stories). Approach 1 is a retrieve-then-generate pipeline using an open-weight Llama 3.1 70B Instruct model with structured reasoning and a self-correction pass. Approach 2 is a hybrid ensemble that combines API-based LLM prompting with transformer representations and a learned calibration layer trained on the development set. On the development set, Approach 2 achieves Spearman \ensuremathρ = 0.7393 (p 10-102) with accuracy 0.8010 (471/588). Approach 1 achieves \ensuremathρ = 0.5187 (p 10-65) with accuracy 0.6032 (561/930). We emphasize that Approach 1 does not exceed RoBERTabase in accuracy (0.6032 vs. 0.6410), but provides stronger rank correlation.</abstract>
<identifier type="citekey">asare-aryal-2026-ai4pc</identifier>
<location>
<url>https://aclanthology.org/2026.semeval-1.326/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>2592</start>
<end>2596</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T AI4PC-Howard University at SemEval-2026 Task 5: Calibrated Hybrid Ensembling and Retrieval-Augmented LLM Reasoning for Narrative Word-Sense Plausibility
%A Asare, Kwaku
%A Aryal, Saurav
%Y Kochmar, Ekaterina
%Y Ghosh, Debanjan
%Y North, Kai
%Y Komachi, Mamoru
%S Proceedings of the 20th International Workshop on Semantic Evaluation (2026)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, USA
%@ 979-8-89176-414-9
%F asare-aryal-2026-ai4pc
%X We present two complementary approaches for rating word-sense plausibility in SemEval-2026 Task 5 (literary homonyms in five-sentence stories). Approach 1 is a retrieve-then-generate pipeline using an open-weight Llama 3.1 70B Instruct model with structured reasoning and a self-correction pass. Approach 2 is a hybrid ensemble that combines API-based LLM prompting with transformer representations and a learned calibration layer trained on the development set. On the development set, Approach 2 achieves Spearman \ensuremathρ = 0.7393 (p 10-102) with accuracy 0.8010 (471/588). Approach 1 achieves \ensuremathρ = 0.5187 (p 10-65) with accuracy 0.6032 (561/930). We emphasize that Approach 1 does not exceed RoBERTabase in accuracy (0.6032 vs. 0.6410), but provides stronger rank correlation.
%U https://aclanthology.org/2026.semeval-1.326/
%P 2592-2596
Markdown (Informal)
[AI4PC-Howard University at SemEval-2026 Task 5: Calibrated Hybrid Ensembling and Retrieval-Augmented LLM Reasoning for Narrative Word-Sense Plausibility](https://aclanthology.org/2026.semeval-1.326/) (Asare & Aryal, SemEval 2026)
ACL