When Individually Calibrated Models Become Collectively Miscalibrated
The paper discusses the phenomenon where individually calibrated models can become collectively miscalibrated in multi-agent settings. This miscalibration occurs even without deliberate coordination when models interact strategically. The study highlights the effectiveness of VCG-based aggregation in aligning incentives and improving performance in various scenarios.
- ▪Individually calibrated predictors can become collectively miscalibrated due to strategic interactions.
- ▪The Price of Anarchy can exceed one when agents have positively correlated beliefs.
- ▪VCG-based aggregation achieves dominant-strategy incentive compatibility and near-optimal performance.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Machine Learning arXiv:2605.18858 (cs) [Submitted on 14 May 2026] Title:When Individually Calibrated Models Become Collectively Miscalibrated Authors:Zhaohui Wang View a PDF of the paper titled When Individually Calibrated Models Become Collectively Miscalibrated, by Zhaohui Wang View PDF HTML (experimental) Abstract:Probabilistic prediction systems often aggregate probability estimates from multiple models into a single decision. A common assumption is that if each model is individually calibrated, the aggregate prediction will also be well calibrated.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
