<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="er-2026-math">
    <titleInfo>
        <title>Math-DB: A Discourse Framework for Mathematical Word Problems to Enhance LLM Reasoning</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Mustafa</namePart>
        <namePart type="given">Erolcan</namePart>
        <namePart type="family">Er</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 Linguistic Annotation Workshop (LAW XX)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Yang</namePart>
            <namePart type="given">Janet</namePart>
            <namePart type="family">Liu</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Luke</namePart>
            <namePart type="family">Gessler</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">979-8-89176-404-0</identifier>
    </relatedItem>
    <abstract>Large Language Models have demonstrated significant progress in solving mathematical word problems through techniques like Chain-of-Thought (CoT) prompting. However, recent research indicates that these models often rely on statistical regularities and surface-level patterns rather than true logical reasoning, leading to performance drops when faced with minor problem perturbations or irrelevant information. In this study, we introduce Math Discourse Bank (Math-DB), a novel discourse framework and annotated dataset designed to enhance LLM reasoning. Inspired by the Penn Discourse TreeBank (PDTB) and mathematics education research, Math-DB defines a hierarchy of discourse senses designed for quantitative reasoning, including categories such as Change, Combine, Compare, and Equalize. We applied this framework to the GSM-Symbolic dataset of 12,500 problems, yielding 47,815 sense-labeled discourse relations over 11,414 successfully-aligned instances (91.3% pipeline yield). Our experiments demonstrate that incorporating Math-DB annotations into CoT prompts consistently improves LLM performance across various difficulty levels.</abstract>
    <identifier type="citekey">er-2026-math</identifier>
    <location>
        <url>https://aclanthology.org/2026.law-main.7/</url>
    </location>
    <part>
        <date>2026-07</date>
        <extent unit="page">
            <start>75</start>
            <end>94</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Math-DB: A Discourse Framework for Mathematical Word Problems to Enhance LLM Reasoning
%A Er, Mustafa Erolcan
%Y Liu, Yang Janet
%Y Gessler, Luke
%S Proceedings of the 20th Linguistic Annotation Workshop (LAW XX)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, USA
%@ 979-8-89176-404-0
%F er-2026-math
%X Large Language Models have demonstrated significant progress in solving mathematical word problems through techniques like Chain-of-Thought (CoT) prompting. However, recent research indicates that these models often rely on statistical regularities and surface-level patterns rather than true logical reasoning, leading to performance drops when faced with minor problem perturbations or irrelevant information. In this study, we introduce Math Discourse Bank (Math-DB), a novel discourse framework and annotated dataset designed to enhance LLM reasoning. Inspired by the Penn Discourse TreeBank (PDTB) and mathematics education research, Math-DB defines a hierarchy of discourse senses designed for quantitative reasoning, including categories such as Change, Combine, Compare, and Equalize. We applied this framework to the GSM-Symbolic dataset of 12,500 problems, yielding 47,815 sense-labeled discourse relations over 11,414 successfully-aligned instances (91.3% pipeline yield). Our experiments demonstrate that incorporating Math-DB annotations into CoT prompts consistently improves LLM performance across various difficulty levels.
%U https://aclanthology.org/2026.law-main.7/
%P 75-94

Markdown (Informal)

[Math-DB: A Discourse Framework for Mathematical Word Problems to Enhance LLM Reasoning](https://aclanthology.org/2026.law-main.7/) (Er, LAW 2026)

• Math-DB: A Discourse Framework for Mathematical Word Problems to Enhance LLM Reasoning (Er, LAW 2026)

ACL

• Mustafa Erolcan Er. 2026. Math-DB: A Discourse Framework for Mathematical Word Problems to Enhance LLM Reasoning. In Proceedings of the 20th Linguistic Annotation Workshop (LAW XX), pages 75–94, San Diego, California, USA. Association for Computational Linguistics.