Valeriy M., PhD, MBA, CQFJun 6
🇧🇷🔥 Brazil on Fire Again – and So Is Its AI Research!
A fresh paper out of São Paulo brings stacking and conformal prediction together in a clever way that boosts predictive performance without splitting off a calibration set — a big win when data is limited.
Show threadValeriy M., PhD, MBA, CQFJun 6

📄 Paper: Stacked Conformal Prediction by Paulo C. Marques
📍 Institution: Insper Institute of Education and Research

🚀 Summary: What’s New?

Show threadValeriy M., PhD, MBA, CQFJun 6
This work proposes a method for conformalizing stacked ensembles of machine learning models — combining the accuracy gains of stacking with the statistical guarantees of conformal prediction.
📌 Key Insight:
Show threadValeriy M., PhD, MBA, CQFJun 6
By using a simple meta-learner (e.g., linear regression) at the top of the stack, you can reuse all available data and avoid holding out a calibration set — all while keeping prediction intervals valid (or approximately valid) under mild assumptions.
🧠 How Conformal Prediction Is Used
Show threadValeriy M., PhD, MBA, CQFJun 6
Conformal prediction is applied at the meta-learner level of the stacked ensemble.
The approach ensures marginal coverage of prediction intervals without the usual data split, by exploiting symmetry in the stacked structure.
Show threadValeriy M., PhD, MBA, CQFJun 6
Even when the ideal symmetry is broken (as in real use), the paper shows approximate validity holds if base models are stable
.
📊 Results (see charts on p.5–6):
On both the California Housing and Ames Housing datasets, the stacked conformal prediction (Stacked CP) method:
Show threadValeriy M., PhD, MBA, CQFJun 6
Achieved comparable or better coverage than Conformalized Quantile Regression (CQR),
Produced shorter median prediction intervals — meaning tighter, more informative uncertainty estimates.
🔧 Open-source code available
Show threadValeriy M., PhD, MBA, CQFJun 6
🧵 This is a powerful contribution for real-world ML where data is tight, calibration matters, and ensemble learning is king. Bravo to Brazilian ML researchers leading the charge!
Show threadValeriy M., PhD, MBA, CQF

#MachineLearning #ConformalPrediction #Stacking #UncertaintyQuantification #BrazilAI #AIResearch #PredictiveInference

https://arxiv.org/abs/2505.12578

Stacked conformal prediction

We consider the conformalization of a stacked ensemble of predictive models, showing that the potentially simple form of the meta-learner at the top of the stack enables a procedure with manageable computational cost that achieves approximate marginal validity without requiring the use of a separate calibration sample. Empirical results indicate that the method compares favorably to a standard inductive alternative.

arXiv.org
Jun 6 at 2:51pmWeb