Metraipolis › archive › dc_ff95923ebbf2

Decide, Don't Generate: Competitive Dimensional ABSA with Jev's Typed Decisions — arXiv:2609.35293

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.35293
License (as archived): CC0-1.0
Archived: 2026-10-07T21:53:40.029Z
SHA-256 of the archived text: 4ba0fdfe0be297d85effeaee84952ea47c15b4648e0a0299ce46f2ac1487144c
Stable URL of this page: https://metraipolis.com/documento/dc_ff95923ebbf2

How to cite

Decide, Don't Generate: Competitive Dimensional ABSA with Jev's Typed Decisions — arXiv:2609.35293 — arXiv (en), https://arxiv.org/abs/2609.35293 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ff95923ebbf2 (Metraipolis archive dc_ff95923ebbf2, sha256 4ba0fdfe0be297d85effeaee84952ea47c15b4648e0a0299ce46f2ac1487144c)

BibTeX

@misc{dcff95923ebbf2,
  title = {Decide, Don't Generate: Competitive Dimensional ABSA with Jev's Typed Decisions — arXiv:2609.35293},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.35293},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ff95923ebbf2; sha256 4ba0fdfe0be297d85effeaee84952ea47c15b4648e0a0299ce46f2ac1487144c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35293. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Aspect-based sentiment analysis (ABSA) has largely turned to text generation. We show that competitive dimensional ABSA does not need it. Using Jev, a frozen model that answers typed questions with rubric scores, label probabilities, and yes/no judgments, we decompose all three tasks of SemEval-2026 Task III Track A into such decisions and align them with the annotation scheme through 488 coefficients fitted on CPU, with no text generation and no backbone tuning. On valence-arousal regression over ten corpora in six languages, the system reaches 1.0645 RMSE, the lowest aggregate error of any participating system. On triplet and quadruplet extraction, it reaches 52.09 and 44.06 continuous F1, above fine-tuned Llama-3.3-70B and GPT-OSS-120B baselines. Analyses and ablations show where the accuracy comes from: supervised calibration roughly halves the raw regression error, exact valence-arousal would add only 4.5 F1 to extraction, and the learned combination of span-boundary evidence, not any single signal, carries the extraction systems.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.35293 — 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 2026-10-07T21:27:08.675Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).