Metraipolis › archive › dc_54ffefc003ef

Learning to Prefer Reliably: Error-Augmented Emotion Preference Optimization with Calibrated Fusion — arXiv:2608.24730

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.24730
License (as archived): CC0-1.0
Archived: 2026-10-18T06:32:13.382Z
SHA-256 of the archived text: 72cb308b68ed9bda9c77f125494ff4475b30b229d115110f62e6f960b21bf955
Stable URL of this page: https://metraipolis.com/documento/dc_54ffefc003ef

How to cite

Learning to Prefer Reliably: Error-Augmented Emotion Preference Optimization with Calibrated Fusion — arXiv:2608.24730 — arXiv (en), https://arxiv.org/abs/2608.24730 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_54ffefc003ef (Metraipolis archive dc_54ffefc003ef, sha256 72cb308b68ed9bda9c77f125494ff4475b30b229d115110f62e6f960b21bf955)

BibTeX

@misc{dc54ffefc003ef,
  title = {Learning to Prefer Reliably: Error-Augmented Emotion Preference Optimization with Calibrated Fusion — arXiv:2608.24730},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.24730},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_54ffefc003ef; sha256 72cb308b68ed9bda9c77f125494ff4475b30b229d115110f62e6f960b21bf955},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.24730. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Emotion preference learning uses pairwise comparisons between candidate descriptions to align multimodal large language models (MLLMs) with human judgments of open-ended emotion descriptions and to train reward models that capture human emotional preferences. However, conventional pairwise supervision is often sparse, typically providing only a single negative description for each positive description, and therefore offers limited coverage of the diverse ways in which an emotion description can be incorrect. In particular, models may be insufficiently exposed to semantically fluent but emotionally inconsistent descriptions. Beyond this data-level limitation, relying on a single MLLM judge introduces a distinct model-level concern: its judgments can be affected by model-specific biases when interpreting fine-grained or ambiguous multimodal emotional cues. To address these limitations, we propose Error-Augmented Preference Optimization (EAPO), a framework for improving the reliability of MLLM-based emotion preference judgment at both the data and model levels. First, we construct an error-augmented dataset by generating multiple controlled and emotion-aware negative descriptions from each preferred description. We then adapt multiple independent MLLM judges to this richer supervision and aggregate their preference margins using margin-calibrated soft fusion, which maps heterogeneous margins to a common scale before aggregation. Experiments on the MER2026-EmoPrefer Challenge dataset and our error-augmented dataset demonstrate that EAPO improves emotion preference prediction and enhances the robustness of MLLM judges when evaluating fluent descriptions that conflict with the video's multimodal emotional evidence. Our code is available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-26 in cs.HC).
Categorie arXiv: Multimedia (cs.MM) ; Human-Computer Interaction (cs.HC).
Pagina del lavoro: https://arxiv.org/abs/2608.24730 — 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 2026-10-18T06:28:16.189Z dalla pagina https://arxiv.org/catchup/cs.HC/2026-08-26?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).