@inproceedings{shrestha-aryal-2026-howard,
title = "{H}oward {U}niversity-{AI}4{PC} at {S}em{E}val-2026 Task 8: Query Reformulation and Dense-Lexical Retrieval Fusion for Multi-Turn Retrieval-Augmented Generation",
author = "Shrestha, Sijan 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-1.320/",
pages = "2533--2539",
ISBN = "979-8-89176-414-9",
abstract = "We present a training-free hybrid retrieve-then-rerank system for multi-turn retrieval-augmented generation, submitted to allthree subtasks of SemEval-2026 Task 8(MTRAGEval): passage retrieval (Task A),generation with reference passages (Task B),and end-to-end RAG (Task C). Our system ad-dresses the core multi-turn challenges{---}non-standalone questions, unanswerable queries,and shifting passage relevance{---}across fourdomain-specific corpora: ClapNQ, Cloud,FiQA, and Govt. Queries are reformulatedthrough LLM-driven rewriting, decompositioninto sub-queries, and Hypothetical DocumentEmbeddings (HyDE). Retrieved candidatesfrom dense vector search (BGE-base-en-v1.5)and BM25 lexical matching are fused via Re-ciprocal Rank Fusion and reranked by a cross-encoder (BGE-reranker-large). Llama-3.3-70B-Instruct generates extractive, context-groundedresponses with built-in abstention for unanswer-able queries. Using only open-source mod-els without fine-tuning, the system achievesnDCG@5 of 0.4098 on Task A (22nd/38), aharmonic mean of 0.7462 on Task B (9th/26),and 0.5796 on Task C (2nd/29), coming within1.1{\%} of the top submission. We attribute thestrong Task C result to the synergy betweenmulti-signal query reformulation and faithfulextractive generation."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="shrestha-aryal-2026-howard">
<titleInfo>
<title>Howard University-AI4PC at SemEval-2026 Task 8: Query Reformulation and Dense-Lexical Retrieval Fusion for Multi-Turn Retrieval-Augmented Generation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sijan</namePart>
<namePart type="family">Shrestha</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 a training-free hybrid retrieve-then-rerank system for multi-turn retrieval-augmented generation, submitted to allthree subtasks of SemEval-2026 Task 8(MTRAGEval): passage retrieval (Task A),generation with reference passages (Task B),and end-to-end RAG (Task C). Our system ad-dresses the core multi-turn challenges—non-standalone questions, unanswerable queries,and shifting passage relevance—across fourdomain-specific corpora: ClapNQ, Cloud,FiQA, and Govt. Queries are reformulatedthrough LLM-driven rewriting, decompositioninto sub-queries, and Hypothetical DocumentEmbeddings (HyDE). Retrieved candidatesfrom dense vector search (BGE-base-en-v1.5)and BM25 lexical matching are fused via Re-ciprocal Rank Fusion and reranked by a cross-encoder (BGE-reranker-large). Llama-3.3-70B-Instruct generates extractive, context-groundedresponses with built-in abstention for unanswer-able queries. Using only open-source mod-els without fine-tuning, the system achievesnDCG@5 of 0.4098 on Task A (22nd/38), aharmonic mean of 0.7462 on Task B (9th/26),and 0.5796 on Task C (2nd/29), coming within1.1% of the top submission. We attribute thestrong Task C result to the synergy betweenmulti-signal query reformulation and faithfulextractive generation.</abstract>
<identifier type="citekey">shrestha-aryal-2026-howard</identifier>
<location>
<url>https://aclanthology.org/2026.semeval-1.320/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>2533</start>
<end>2539</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Howard University-AI4PC at SemEval-2026 Task 8: Query Reformulation and Dense-Lexical Retrieval Fusion for Multi-Turn Retrieval-Augmented Generation
%A Shrestha, Sijan
%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 shrestha-aryal-2026-howard
%X We present a training-free hybrid retrieve-then-rerank system for multi-turn retrieval-augmented generation, submitted to allthree subtasks of SemEval-2026 Task 8(MTRAGEval): passage retrieval (Task A),generation with reference passages (Task B),and end-to-end RAG (Task C). Our system ad-dresses the core multi-turn challenges—non-standalone questions, unanswerable queries,and shifting passage relevance—across fourdomain-specific corpora: ClapNQ, Cloud,FiQA, and Govt. Queries are reformulatedthrough LLM-driven rewriting, decompositioninto sub-queries, and Hypothetical DocumentEmbeddings (HyDE). Retrieved candidatesfrom dense vector search (BGE-base-en-v1.5)and BM25 lexical matching are fused via Re-ciprocal Rank Fusion and reranked by a cross-encoder (BGE-reranker-large). Llama-3.3-70B-Instruct generates extractive, context-groundedresponses with built-in abstention for unanswer-able queries. Using only open-source mod-els without fine-tuning, the system achievesnDCG@5 of 0.4098 on Task A (22nd/38), aharmonic mean of 0.7462 on Task B (9th/26),and 0.5796 on Task C (2nd/29), coming within1.1% of the top submission. We attribute thestrong Task C result to the synergy betweenmulti-signal query reformulation and faithfulextractive generation.
%U https://aclanthology.org/2026.semeval-1.320/
%P 2533-2539
Markdown (Informal)
[Howard University-AI4PC at SemEval-2026 Task 8: Query Reformulation and Dense-Lexical Retrieval Fusion for Multi-Turn Retrieval-Augmented Generation](https://aclanthology.org/2026.semeval-1.320/) (Shrestha & Aryal, SemEval 2026)
ACL