<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="liu-etal-2026-unified">
    <titleInfo>
        <title>Unified Fragment Growing, Linking, and Merging with Energy-Aware Pre-training and Reward-Ranked Alignment</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Xuefeng</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Songhao</namePart>
        <namePart type="family">Jiang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Tinson</namePart>
        <namePart type="family">Xu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Qinan</namePart>
        <namePart type="family">Huang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Rick</namePart>
        <namePart type="given">L</namePart>
        <namePart type="family">Stevens</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
    </relatedItem>
    <abstract>Fragment-based drug discovery optimizes small molecular hits through three recurring transformations: fragment growing, linking, and merging. Existing generative methods, however, usually specialize in a single transformation and have focused primarily on linker design. We introduce FragmentGPT, a unified chemical language model that casts all three operations as conditional molecular generation. To build this capability, we construct a chemically-aware pre-training corpus using energy-guided bond cleavage, together with overlap-based examples that teach the model to connect separated fragments and consolidate shared substructures. We further introduce Reward-Ranked Alignment with Expert Exploration (RAE), which combines expert-guided candidate generation, Pareto- and composite-reward-based selection, and supervised fine-tuning for multi-objective molecular optimization. Conditioned on one or more fragments, FragmentGPT can expand a scaffold, connect distinct subunits, or merge overlapping structures into a complete molecule. Experiments on RTCB and SARS-CoV-2 main protease (3CLPro), together with ablation studies, show high chemical validity and strong aggregate performance across multiple drug-design objectives.</abstract>
    <identifier type="citekey">liu-etal-2026-unified</identifier>
    <identifier type="doi">10.18653/2026.findings-emnlp.1524</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.1524/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>32439</start>
            <end>32452</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Unified Fragment Growing, Linking, and Merging with Energy-Aware Pre-training and Reward-Ranked Alignment
%A Liu, Xuefeng
%A Jiang, Songhao
%A Xu, Tinson
%A Huang, Qinan
%A Stevens, Rick L.
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F liu-etal-2026-unified
%X Fragment-based drug discovery optimizes small molecular hits through three recurring transformations: fragment growing, linking, and merging. Existing generative methods, however, usually specialize in a single transformation and have focused primarily on linker design. We introduce FragmentGPT, a unified chemical language model that casts all three operations as conditional molecular generation. To build this capability, we construct a chemically-aware pre-training corpus using energy-guided bond cleavage, together with overlap-based examples that teach the model to connect separated fragments and consolidate shared substructures. We further introduce Reward-Ranked Alignment with Expert Exploration (RAE), which combines expert-guided candidate generation, Pareto- and composite-reward-based selection, and supervised fine-tuning for multi-objective molecular optimization. Conditioned on one or more fragments, FragmentGPT can expand a scaffold, connect distinct subunits, or merge overlapping structures into a complete molecule. Experiments on RTCB and SARS-CoV-2 main protease (3CLPro), together with ablation studies, show high chemical validity and strong aggregate performance across multiple drug-design objectives.
%R 10.18653/2026.findings-emnlp.1524
%U https://aclanthology.org/2026.findings-emnlp.1524/
%U https://doi.org/10.18653/2026.findings-emnlp.1524
%P 32439-32452

Markdown (Informal)

[Unified Fragment Growing, Linking, and Merging with Energy-Aware Pre-training and Reward-Ranked Alignment](https://aclanthology.org/2026.findings-emnlp.1524/) (Liu et al., Findings 2026)

• Unified Fragment Growing, Linking, and Merging with Energy-Aware Pre-training and Reward-Ranked Alignment (Liu et al., Findings 2026)

ACL

• Xuefeng Liu, Songhao Jiang, Tinson Xu, Qinan Huang, and Rick L. Stevens. 2026. Unified Fragment Growing, Linking, and Merging with Energy-Aware Pre-training and Reward-Ranked Alignment. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 32439–32452, Budapest, Hungary. Association for Computational Linguistics.