@inproceedings{pfister-etal-2026-assembly,
title = "Some Assembly Required: Pointer-Decoded {S}eq2{S}eq Models for {G}erman Without a Native {S}eq2{S}eq Backbone",
author = {Pfister, Jan and
V{\"o}lker, Tom and
Hotho, Andreas},
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.2464/",
pages = "53012--53027",
ISBN = "979-8-89176-462-0",
abstract = {Pointer-decoded Seq2Seq transformers remain a strong recipe for aspect-based sentiment analysis (ABSA), but native German Seq2Seq backbones remain scarce, and current pretraining trends favor decoder-only architectures. Existing assembly recipes disagree on which pairings work. We therefore study models assembled from independently pretrained encoder and decoder LMs, primarily on German ABSA, where fixed pointer outputs expose both tuple accuracy and grammar violations. All nine backbone combinations train under both separate and parameter-sharing merged cross-attention, with separate attention slightly stronger overall. Depth and scale sweeps show that additional capacity pays off more on the encoder side for these pointer-decoded tasks, and our strongest two-stack assembly trails mBART but outperforms mT5-small on both German datasets. Motivated by decoder-only pretraining, we replace the encoder and decoder with one backbone that encodes the source and autoregressively produces pointer queries. Bidirectional source visibility and scale matter more consistently than pointer-query refinement: a 120M LL{\"a}Mmlein model exceeds both same-family two-stack means with roughly half the parameters, and a 1B ModernGBERT reaches mBART-level performance.}
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="pfister-etal-2026-assembly">
<titleInfo>
<title>Some Assembly Required: Pointer-Decoded Seq2Seq Models for German Without a Native Seq2Seq Backbone</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jan</namePart>
<namePart type="family">Pfister</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tom</namePart>
<namePart type="family">Völker</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andreas</namePart>
<namePart type="family">Hotho</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>Pointer-decoded Seq2Seq transformers remain a strong recipe for aspect-based sentiment analysis (ABSA), but native German Seq2Seq backbones remain scarce, and current pretraining trends favor decoder-only architectures. Existing assembly recipes disagree on which pairings work. We therefore study models assembled from independently pretrained encoder and decoder LMs, primarily on German ABSA, where fixed pointer outputs expose both tuple accuracy and grammar violations. All nine backbone combinations train under both separate and parameter-sharing merged cross-attention, with separate attention slightly stronger overall. Depth and scale sweeps show that additional capacity pays off more on the encoder side for these pointer-decoded tasks, and our strongest two-stack assembly trails mBART but outperforms mT5-small on both German datasets. Motivated by decoder-only pretraining, we replace the encoder and decoder with one backbone that encodes the source and autoregressively produces pointer queries. Bidirectional source visibility and scale matter more consistently than pointer-query refinement: a 120M LLäMmlein model exceeds both same-family two-stack means with roughly half the parameters, and a 1B ModernGBERT reaches mBART-level performance.</abstract>
<identifier type="citekey">pfister-etal-2026-assembly</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2464/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>53012</start>
<end>53027</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Some Assembly Required: Pointer-Decoded Seq2Seq Models for German Without a Native Seq2Seq Backbone
%A Pfister, Jan
%A Völker, Tom
%A Hotho, Andreas
%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 pfister-etal-2026-assembly
%X Pointer-decoded Seq2Seq transformers remain a strong recipe for aspect-based sentiment analysis (ABSA), but native German Seq2Seq backbones remain scarce, and current pretraining trends favor decoder-only architectures. Existing assembly recipes disagree on which pairings work. We therefore study models assembled from independently pretrained encoder and decoder LMs, primarily on German ABSA, where fixed pointer outputs expose both tuple accuracy and grammar violations. All nine backbone combinations train under both separate and parameter-sharing merged cross-attention, with separate attention slightly stronger overall. Depth and scale sweeps show that additional capacity pays off more on the encoder side for these pointer-decoded tasks, and our strongest two-stack assembly trails mBART but outperforms mT5-small on both German datasets. Motivated by decoder-only pretraining, we replace the encoder and decoder with one backbone that encodes the source and autoregressively produces pointer queries. Bidirectional source visibility and scale matter more consistently than pointer-query refinement: a 120M LLäMmlein model exceeds both same-family two-stack means with roughly half the parameters, and a 1B ModernGBERT reaches mBART-level performance.
%U https://aclanthology.org/2026.findings-emnlp.2464/
%P 53012-53027
Markdown (Informal)
[Some Assembly Required: Pointer-Decoded Seq2Seq Models for German Without a Native Seq2Seq Backbone](https://aclanthology.org/2026.findings-emnlp.2464/) (Pfister et al., Findings 2026)
ACL