@inproceedings{matta-etal-2025-optimizing,
title = "Optimizing Cost-Efficiency with {LLM}-Generated Training Data for Conversational Semantic Frame Analysis",
author = "Matta, Shiho and
Huang, Yin Jou and
Cheng, Fei and
Kiyomaru, Hirokazu and
Murawaki, Yugo",
editor = "Kazantseva, Anna and
Szpakowicz, Stan and
Degaetano-Ortlieb, Stefania and
Bizzoni, Yuri and
Pagel, Janis",
booktitle = "Proceedings of the 9th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature (LaTeCH-CLfL 2025)",
month = may,
year = "2025",
address = "Albuquerque, New Mexico",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.latechclfl-1.21/",
doi = "10.18653/v1/2025.latechclfl-1.21",
pages = "238--251",
ISBN = "979-8-89176-241-1",
abstract = "Recent studies have shown that few-shot learning enables large language models (LLMs) to generate training data for supervised models at a low cost. However, for complex tasks, the quality of LLM-generated data often falls short compared to human-labeled data. This presents a critical challenge: how should one balance the trade-off between the higher quality but more expensive human-annotated data and the lower quality yet significantly cheaper LLM-generated data? In this paper, we tackle this question for a demanding task: conversational semantic frame analysis (SFA). To address this, we propose a novel method for synthesizing training data tailored to this complex task. Through experiments conducted across a wide range of budget levels, we find that smaller budgets favor a higher reliance on LLM-generated data to achieve optimal cost-efficiency."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="matta-etal-2025-optimizing">
<titleInfo>
<title>Optimizing Cost-Efficiency with LLM-Generated Training Data for Conversational Semantic Frame Analysis</title>
</titleInfo>
<name type="personal">
<namePart type="given">Shiho</namePart>
<namePart type="family">Matta</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yin</namePart>
<namePart type="given">Jou</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fei</namePart>
<namePart type="family">Cheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hirokazu</namePart>
<namePart type="family">Kiyomaru</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yugo</namePart>
<namePart type="family">Murawaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 9th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature (LaTeCH-CLfL 2025)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anna</namePart>
<namePart type="family">Kazantseva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Stan</namePart>
<namePart type="family">Szpakowicz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Stefania</namePart>
<namePart type="family">Degaetano-Ortlieb</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuri</namePart>
<namePart type="family">Bizzoni</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Janis</namePart>
<namePart type="family">Pagel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Albuquerque, New Mexico</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-241-1</identifier>
</relatedItem>
<abstract>Recent studies have shown that few-shot learning enables large language models (LLMs) to generate training data for supervised models at a low cost. However, for complex tasks, the quality of LLM-generated data often falls short compared to human-labeled data. This presents a critical challenge: how should one balance the trade-off between the higher quality but more expensive human-annotated data and the lower quality yet significantly cheaper LLM-generated data? In this paper, we tackle this question for a demanding task: conversational semantic frame analysis (SFA). To address this, we propose a novel method for synthesizing training data tailored to this complex task. Through experiments conducted across a wide range of budget levels, we find that smaller budgets favor a higher reliance on LLM-generated data to achieve optimal cost-efficiency.</abstract>
<identifier type="citekey">matta-etal-2025-optimizing</identifier>
<identifier type="doi">10.18653/v1/2025.latechclfl-1.21</identifier>
<location>
<url>https://aclanthology.org/2025.latechclfl-1.21/</url>
</location>
<part>
<date>2025-05</date>
<extent unit="page">
<start>238</start>
<end>251</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Optimizing Cost-Efficiency with LLM-Generated Training Data for Conversational Semantic Frame Analysis
%A Matta, Shiho
%A Huang, Yin Jou
%A Cheng, Fei
%A Kiyomaru, Hirokazu
%A Murawaki, Yugo
%Y Kazantseva, Anna
%Y Szpakowicz, Stan
%Y Degaetano-Ortlieb, Stefania
%Y Bizzoni, Yuri
%Y Pagel, Janis
%S Proceedings of the 9th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature (LaTeCH-CLfL 2025)
%D 2025
%8 May
%I Association for Computational Linguistics
%C Albuquerque, New Mexico
%@ 979-8-89176-241-1
%F matta-etal-2025-optimizing
%X Recent studies have shown that few-shot learning enables large language models (LLMs) to generate training data for supervised models at a low cost. However, for complex tasks, the quality of LLM-generated data often falls short compared to human-labeled data. This presents a critical challenge: how should one balance the trade-off between the higher quality but more expensive human-annotated data and the lower quality yet significantly cheaper LLM-generated data? In this paper, we tackle this question for a demanding task: conversational semantic frame analysis (SFA). To address this, we propose a novel method for synthesizing training data tailored to this complex task. Through experiments conducted across a wide range of budget levels, we find that smaller budgets favor a higher reliance on LLM-generated data to achieve optimal cost-efficiency.
%R 10.18653/v1/2025.latechclfl-1.21
%U https://aclanthology.org/2025.latechclfl-1.21/
%U https://doi.org/10.18653/v1/2025.latechclfl-1.21
%P 238-251
Markdown (Informal)
[Optimizing Cost-Efficiency with LLM-Generated Training Data for Conversational Semantic Frame Analysis](https://aclanthology.org/2025.latechclfl-1.21/) (Matta et al., LaTeCHCLfL 2025)
ACL