<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="de-lange-etal-2026-unified">
    <titleInfo>
        <title>Unified Work Embeddings: Contrastive Learning of a Bidirectional Multi-task Ranker</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Matthias</namePart>
        <namePart type="family">De Lange</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Warre</namePart>
        <namePart type="family">Veys</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jens-Joris</namePart>
        <namePart type="family">Decorte</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jeroen</namePart>
        <namePart type="family">Van Hautte</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Hung-Yi</namePart>
            <namePart type="family">Lee</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sunipa</namePart>
            <namePart type="family">Dev</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jessy</namePart>
            <namePart type="family">Li</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Desmond</namePart>
            <namePart type="family">Elliott</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-459-0</identifier>
    </relatedItem>
    <abstract>Applications in labor market intelligence demand specialized NLP systems for a wide range of tasks, characterized by extreme multi-label target spaces, strict latency constraints, and multiple text modalities such as skills and job titles. These constraints have led to isolated, task-specific developments in the field, with models and benchmarks focused on single prediction tasks. Exploiting the shared structure of work-related data, we propose a unifying framework, combining a wide range of tasks in a multi-task ranking benchmark, and a flexible architecture tackling text-driven work tasks with a single model. The benchmark, WorkBench, is the first unified evaluation suite spanning six work-related tasks formulated explicitly as ranking problems, curated from real-world ontologies and human-annotated resources. WorkBench enables cross-task analysis, where we find significant positive cross-task transfer. This insight leads to Unified Work Embeddings (UWE), a task-agnostic bi-encoder that exploits our training-data structure with a many-to-many InfoNCE objective, and leverages token-level embeddings with task-agnostic soft late interaction. UWE demonstrates zero-shot ranking performance on unseen target spaces in the work domain, and enables low-latency inference with two orders of magnitude fewer parameters than best-performing generalist models that are not tailored to the work domain (Qwen3-8B), with +4.4 MAP improvement.</abstract>
    <identifier type="citekey">de-lange-etal-2026-unified</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.442/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>10196</start>
            <end>10213</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Unified Work Embeddings: Contrastive Learning of a Bidirectional Multi-task Ranker
%A De Lange, Matthias
%A Veys, Warre
%A Decorte, Jens-Joris
%A Van Hautte, Jeroen
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F de-lange-etal-2026-unified
%X Applications in labor market intelligence demand specialized NLP systems for a wide range of tasks, characterized by extreme multi-label target spaces, strict latency constraints, and multiple text modalities such as skills and job titles. These constraints have led to isolated, task-specific developments in the field, with models and benchmarks focused on single prediction tasks. Exploiting the shared structure of work-related data, we propose a unifying framework, combining a wide range of tasks in a multi-task ranking benchmark, and a flexible architecture tackling text-driven work tasks with a single model. The benchmark, WorkBench, is the first unified evaluation suite spanning six work-related tasks formulated explicitly as ranking problems, curated from real-world ontologies and human-annotated resources. WorkBench enables cross-task analysis, where we find significant positive cross-task transfer. This insight leads to Unified Work Embeddings (UWE), a task-agnostic bi-encoder that exploits our training-data structure with a many-to-many InfoNCE objective, and leverages token-level embeddings with task-agnostic soft late interaction. UWE demonstrates zero-shot ranking performance on unseen target spaces in the work domain, and enables low-latency inference with two orders of magnitude fewer parameters than best-performing generalist models that are not tailored to the work domain (Qwen3-8B), with +4.4 MAP improvement.
%U https://aclanthology.org/2026.emnlp-main.442/
%P 10196-10213

Markdown (Informal)

[Unified Work Embeddings: Contrastive Learning of a Bidirectional Multi-task Ranker](https://aclanthology.org/2026.emnlp-main.442/) (De Lange et al., EMNLP 2026)

• Unified Work Embeddings: Contrastive Learning of a Bidirectional Multi-task Ranker (De Lange et al., EMNLP 2026)

ACL

• Matthias De Lange, Warre Veys, Jens-Joris Decorte, and Jeroen Van Hautte. 2026. Unified Work Embeddings: Contrastive Learning of a Bidirectional Multi-task Ranker. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 10196–10213, Budapest, Hungary. Association for Computational Linguistics.