Metraipolis › archive › dc_a19d97a8fda4

Dynamic ordering in sequence generation in Transformers: experiments on low-resource Tagging scenarios — arXiv:2610.21862

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.21862
License (as archived)
CC0-1.0
Archived
2026-10-16T10:01:54.612Z
SHA-256 of the archived text
82b4f2eb78b6603eedbd7258b54cef3770bebf0e122b485cb07bcd8ab3ce4380
Stable URL of this page
https://metraipolis.com/documento/dc_a19d97a8fda4

How to cite

Dynamic ordering in sequence generation in Transformers: experiments on low-resource Tagging scenarios — arXiv:2610.21862 — arXiv (en), https://arxiv.org/abs/2610.21862 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a19d97a8fda4 (Metraipolis archive dc_a19d97a8fda4, sha256 82b4f2eb78b6603eedbd7258b54cef3770bebf0e122b485cb07bcd8ab3ce4380)
BibTeX
@misc{dca19d97a8fda4,
  title = {Dynamic ordering in sequence generation in Transformers: experiments on low-resource Tagging scenarios — arXiv:2610.21862},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.21862},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a19d97a8fda4; sha256 82b4f2eb78b6603eedbd7258b54cef3770bebf0e122b485cb07bcd8ab3ce4380},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21862. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

This paper introduces TagInsert, a Transformer-based model that challenges the common left-to-right (L2R) generation paradigm in Transformers, for sequence labelling tasks where dependencies may not naturally align with a L2R order. TagInsert lifts the L2R constraint by removing the decoder-side attention mask and enabling the model to decide a preferred order of tag generation at inference time. We experiment on the common Part-of-Speech tagging and the more complex Combinatory Categorial Grammar (CCG) supertagging tasks on multiple languages and low-resource data settings. We show that, although encoder-only models like BERT remain strong performers for tagging tasks, TagInsert frequently outperforms a L2R encoder-decoder model, achieves results comparable to encoder-only models under data-scarce conditions, and shows better results on the long-tail of rare categories. Our analyses show that TagInsert alleviates error propagation typical of autoregressive models by learning to defer difficult predictions to later stages in the decoding process.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.21862 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-16T08:44:47.679Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).