<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="shu-zhou-2026-multi">
    <titleInfo>
        <title>Multi-Task Learning Trade-offs in Vision–Language Models for Ancient Chinese OCR: An Empirical Analysis of Parameter-Efficient Adaptation</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Yuhan</namePart>
        <namePart type="family">Shu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Huizi</namePart>
        <namePart type="family">Zhou</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-05</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the Fourth Workshop on Language Technologies for Historical and Ancient Languages (LT4HALA 2026) @ LREC 2026</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Rachele</namePart>
            <namePart type="family">Sprugnoli</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Marco</namePart>
            <namePart type="family">Passarotti</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>ELRA Language Resources Association (ELRA)</publisher>
            <place>
                <placeTerm type="text">Palma, Mallorca (Spain)</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>This study evaluates the efficacy of multi-task adaptation in large-scale vision–language models (VLMs), specifically Qwen2.5-VL, for the simultaneous recognition and structural parsing of historical Chinese documents within the EvaHan2026 benchmark. Utilizing a parameter-efficient fine-tuning (PEFT) strategy via LoRA (rank 64), our framework demonstrates superior performance in layout analysis (Task B), achieving an mAP of 0.2802—a 39.6% improvement over the competitive baseline—and a Macro F1 of 0.3609. Conversely, a pronounced performance-utility trade-off is observed in printed OCR (Task A), where the character error rate (CER) escalates from 0.0618 to 0.1100 (+78% relative). This divergence highlights a critical catastrophic forgetting effect induced by gradient interference during multi-task optimization. While handwritten OCR (Task C) remains relatively stable (CER of 0.0963), our findings suggest that although unified VLM architectures excel at high-level structural detection, they encounter significant parameter capacity bottlenecks when concurrently optimizing fine-grained character-level transcription. This analysis highlights the optimization challenges when balancing spatial detection and character recognition in a unified framework.</abstract>
    <identifier type="citekey">shu-zhou-2026-multi</identifier>
    <identifier type="doi">10.63317/5hvjskashyjv</identifier>
    <location>
        <url>https://aclanthology.org/2026.lt4hala-1.34/</url>
    </location>
    <part>
        <date>2026-05</date>
        <extent unit="page">
            <start>330</start>
            <end>338</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Multi-Task Learning Trade-offs in Vision–Language Models for Ancient Chinese OCR: An Empirical Analysis of Parameter-Efficient Adaptation
%A Shu, Yuhan
%A Zhou, Huizi
%Y Sprugnoli, Rachele
%Y Passarotti, Marco
%S Proceedings of the Fourth Workshop on Language Technologies for Historical and Ancient Languages (LT4HALA 2026) @ LREC 2026
%D 2026
%8 May
%I ELRA Language Resources Association (ELRA)
%C Palma, Mallorca (Spain)
%F shu-zhou-2026-multi
%X This study evaluates the efficacy of multi-task adaptation in large-scale vision–language models (VLMs), specifically Qwen2.5-VL, for the simultaneous recognition and structural parsing of historical Chinese documents within the EvaHan2026 benchmark. Utilizing a parameter-efficient fine-tuning (PEFT) strategy via LoRA (rank 64), our framework demonstrates superior performance in layout analysis (Task B), achieving an mAP of 0.2802—a 39.6% improvement over the competitive baseline—and a Macro F1 of 0.3609. Conversely, a pronounced performance-utility trade-off is observed in printed OCR (Task A), where the character error rate (CER) escalates from 0.0618 to 0.1100 (+78% relative). This divergence highlights a critical catastrophic forgetting effect induced by gradient interference during multi-task optimization. While handwritten OCR (Task C) remains relatively stable (CER of 0.0963), our findings suggest that although unified VLM architectures excel at high-level structural detection, they encounter significant parameter capacity bottlenecks when concurrently optimizing fine-grained character-level transcription. This analysis highlights the optimization challenges when balancing spatial detection and character recognition in a unified framework.
%R 10.63317/5hvjskashyjv
%U https://aclanthology.org/2026.lt4hala-1.34/
%U https://doi.org/10.63317/5hvjskashyjv
%P 330-338

Markdown (Informal)

[Multi-Task Learning Trade-offs in Vision–Language Models for Ancient Chinese OCR: An Empirical Analysis of Parameter-Efficient Adaptation](https://aclanthology.org/2026.lt4hala-1.34/) (Shu & Zhou, LT4HALA 2026)

• Multi-Task Learning Trade-offs in Vision–Language Models for Ancient Chinese OCR: An Empirical Analysis of Parameter-Efficient Adaptation (Shu & Zhou, LT4HALA 2026)

ACL

• Yuhan Shu and Huizi Zhou. 2026. Multi-Task Learning Trade-offs in Vision–Language Models for Ancient Chinese OCR: An Empirical Analysis of Parameter-Efficient Adaptation. In Proceedings of the Fourth Workshop on Language Technologies for Historical and Ancient Languages (LT4HALA 2026) @ LREC 2026, pages 330–338, Palma, Mallorca (Spain). ELRA Language Resources Association (ELRA).