Metraipolis › archive › dc_808bbc1dd2b0

UO-FIE: Combining Exact-Label Supervision with Graded Utility for Factivity Inference — arXiv:2609.28605

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.28605
License (as archived)
CC0-1.0
Archived
2026-10-08T04:27:03.498Z
SHA-256 of the archived text
62706eb89c4fbb746e1e359ca9af82bc3174efef80d004479f301f4a417729e4
Stable URL of this page
https://metraipolis.com/documento/dc_808bbc1dd2b0

How to cite

UO-FIE: Combining Exact-Label Supervision with Graded Utility for Factivity Inference — arXiv:2609.28605 — arXiv (en), https://arxiv.org/abs/2609.28605 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_808bbc1dd2b0 (Metraipolis archive dc_808bbc1dd2b0, sha256 62706eb89c4fbb746e1e359ca9af82bc3174efef80d004479f301f4a417729e4)
BibTeX
@misc{dc808bbc1dd2b0,
  title = {UO-FIE: Combining Exact-Label Supervision with Graded Utility for Factivity Inference — arXiv:2609.28605},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.28605},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_808bbc1dd2b0; sha256 62706eb89c4fbb746e1e359ca9af82bc3174efef80d004479f301f4a417729e4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.28605. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

The Factivity Inference Evaluation 2026 (FIE2026) classifies Chinese context-hypothesis pairs into nine ordered factivity intervals. Its evaluation metric rewards both exact predictions and proximity to the correct interval, while 64.1% of the 566 training examples belong to a single class. In preliminary experiments, several mDeBERTa classification models predominantly predict the dominant class, whereas a Huber-regression baseline produces more predictions near the correct interval but fewer exact matches. We introduce Utility-Oriented Factivity Inference (UO-FIE), a parameter-efficient system that combines exact-label supervision with graded utility. UO-FIE predicts a distribution over the nine classes and combines hard-label supervision, utility-based soft targets, scheduled class weights, and an ordinal loss. We evaluate expected-utility decoding in controlled comparisons and use ordinal calibration selected on out-of-fold predictions for the submitted system. Based on Qwen3.5-9B with LoRA, UO-FIE ranks first in the fine-tuning track with a macro utility of 0.8316. A separate prompt-based ensemble ranks third in the non-fine-tuning track with a macro utility of 0.8450.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-25 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.28605 — 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 2026-10-08T03:50:57.938Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).