Skip to content
AI News HubLIVE
Original source2 min read

What a Calibrated Probability Hides: Belief-Level Post-Training for Decision Models

Summary

The paper introduces Belief-Grouped Policy Optimization (BGPO), which separates disagreement among confident groups from uncertainty within a shared belief, and achieves lower Jensen-Shannon divergence on several benchmarks.

SourcearXiv Computer VisionAuthor: Xinye Li, Zirui Song, Huaxing Liu, Xiang Wang, Shuai Li, Zixiang Xu, Yinjun Liu, Zhewen Zhang, Xiaojun Chang, Xiuying Chen
What a Calibrated Probability Hides: Belief-Level Post-Training for Decision Models
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 18 Nov 2026]

Title:What a Calibrated Probability Hides: Belief-Level Post-Training for Decision Models

View a PDF of the paper titled What a Calibrated Probability Hides: Belief-Level Post-Training for Decision Models, by Xinye Li and 9 other authors

View PDF HTML (experimental)

Abstract:A calibrated decision model can predict how often each answer is chosen. The same prediction can reflect disagreement among people who are confident in their choices or uncertainty among people who share a belief. Disagreement calls for identifying whose perspective applies, while uncertainty calls for gathering more evidence. Standard post-training scores a single prediction against one observed choice at a time, learning the pooled distribution without distinguishing these two sources of uncertainty. We introduce Belief-Grouped Policy Optimization (BGPO). The policy proposes a list of beliefs, estimates their prevalence, and predicts a choice distribution for each. During training, BGPO uses Bayes' rule to estimate how likely each belief was to generate an observed choice. These estimates weight the choice's contribution to each belief's log-score reward. To discover rare beliefs, the policy also proposes lists after seeing a few observed choices and retains them only if they predict held-out choices well. At inference, the policy uses only the question and combines the belief-specific distributions by prevalence to produce a calibrated aggregate prediction. On benchmarks covering opinions, conversation safety, emotions elicited by artwork, and photo aesthetics, BGPO achieves the lowest conditional, worst-group, and overall Jensen-Shannon divergence across model initializations. It improves conditional divergence on both text and image tasks. On a synthetic image task with known beliefs, BGPO recovers the underlying groups more accurately than baselines trained with belief labels.

Subjects:

Computer Vision and Pattern Recognition (cs.CV)

Cite as: arXiv:2611.25168 [cs.CV]

(or arXiv:2611.25168v1 [cs.CV] for this version)

https://doi.org/10.48550/arXiv.2611.25168

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Zirui Song [view email] [v1] Wed, 18 Nov 2026 20:48:01 UTC (1,389 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled What a Calibrated Probability Hides: Belief-Level Post-Training for Decision Models, by Xinye Li and 9 other authors

View PDF

HTML (experimental)

TeX Source

view license

Additional Features

Audio Summary

Current browse context:

cs.CV

new | recent | 2026-11

Change to browse by:

cs

References & Citations

SciX

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • Standard post-training learns a pooled distribution and cannot tell population disagreement from within-belief uncertainty.
  • BGPO proposes belief hypotheses, estimates their prevalence, and uses Bayes' rule to weight each observed choice's contribution to belief-specific rewards.
  • Across opinion, safety, emotion, and aesthetics benchmarks, BGPO attains the lowest conditional, worst-group, and overall JS divergence.

Highlights and analysis are generated automatically and may contain errors. Check the original source.