# Coupled Training with Privileged Information and Unlabeled Data

arXiv: https://arxiv.org/abs/2605.23268 · HTML: https://mlsanity.com/paper/2605.23268/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 13 of its 32 references that are in the map.
- First posted on arXiv 22 May 2026.

## Builds on (most cited first, 13 of 13)

- [WILDS: A Benchmark of in-the-Wild Distribution Shifts](https://mlsanity.com/paper/2012.07421/) (2020)
- [A Comprehensive Survey on Transfer Learning](https://mlsanity.com/paper/1911.02685/) (2019)
- [Temporal Ensembling for Semi-Supervised Learning](https://mlsanity.com/paper/1610.02242/) (2016)
- [Learning from Noisy Labels with Deep Neural Networks: A Survey](https://mlsanity.com/paper/2007.08199/) (2020)
- [Snorkel: Rapid Training Data Creation with Weak Supervision](https://mlsanity.com/paper/1711.10160/) (2017)
- [Data Programming: Creating Large Training Sets, Quickly](https://mlsanity.com/paper/1605.07723/) (2016)
- [Semi-supervised Inference: General Theory and Estimation of Means](https://mlsanity.com/paper/1606.07268/) (2016)
- [Efficient and Adaptive Linear Regression in Semi-Supervised Settings](https://mlsanity.com/paper/1701.04889/) (2017)
- [Cooperative learning for multiview analysis](https://mlsanity.com/paper/2112.12337/) (2021)
- [Strength from Weakness: Fast Learning Using Weak Supervision](https://mlsanity.com/paper/2002.08483/) (2020)
- [Pseudo-Labeling for Kernel Ridge Regression under Covariate Shift](https://mlsanity.com/paper/2302.10160/) (2023)
- [Surrogate Assisted Semi-supervised Inference for High Dimensional Risk Prediction](https://mlsanity.com/paper/2105.01264/) (2021)
- [Convergence of alternating minimisation algorithms for dictionary learning](https://mlsanity.com/paper/2304.01768/) (2023)

## Built on by (most cited first, 1 of 1)

- [Weak-to-Strong Learning in Decision Making](https://mlsanity.com/paper/2607.18467/) (2026)

## Abstract (arXiv)

In many prediction problems, we have extra information during training (for example, measurements that are expensive or slow to collect) that will not be available when the model is deployed. A common strategy is to first train a model that uses all training information, then use its predictions on unlabeled examples to train a second model that only uses the inputs available at test time. However, when the extra training-only information is weak or noisy, this Two-Stage approach can mislead the deployment model and even hurt accuracy. We propose a joint training method that learns the two models together, so the deployment model can benefit from the extra information only when it actually helps, instead of inheriting its mistakes. We provide guarantees that describe when joint training improves prediction accuracy and analyze a simple alternating training algorithm for large, high-dimensional models. Experiments on synthetic data and real-world prediction tasks show that our approach avoids these failures and robustly outperforms standard Two-Stage baselines.