# Deep Learning under Privileged Information Using Heteroscedastic Dropout

arXiv: https://arxiv.org/abs/1805.11614 · HTML: https://mlsanity.com/paper/1805.11614/

## Key facts

- Cited by 27 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 May 2018.

## Built on by (most cited first, 27 of 27)

- [Object DGCNN: 3D Object Detection using Dynamic Graphs](https://mlsanity.com/paper/2110.06923/) (2021)
- [Human Pose Estimation in Extremely Low-Light Conditions](https://mlsanity.com/paper/2303.15410/) (2023)
- [Trust, but Verify: Cross-Modality Fusion for HD Map Change Detection](https://mlsanity.com/paper/2212.07312/) (2022)
- [Transfer and Marginalize: Explaining Away Label Noise with Privileged Information](https://mlsanity.com/paper/2202.09244/) (2022)
- [Multimodal Machine Learning in Image-Based and Clinical Biomedicine: Survey and Prospects](https://mlsanity.com/paper/2311.02332/) (2023)
- [Make Me a BNN: A Simple Strategy for Estimating Bayesian Uncertainty from Pre-trained Models](https://mlsanity.com/paper/2312.15297/) (2023)
- [A Super-human Vision-based Reinforcement Learning Agent for Autonomous Racing in Gran Turismo](https://mlsanity.com/paper/2406.12563/) (2024)
- [Visible-Infrared Person Re-Identification Using Privileged Intermediate Information](https://mlsanity.com/paper/2209.09348/) (2022)
- [When does Privileged Information Explain Away Label Noise?](https://mlsanity.com/paper/2303.01806/) (2023)
- [HalluciDet: Hallucinating RGB Modality for Person Detection Through Privileged Information](https://mlsanity.com/paper/2310.04662/) (2023)
- [Adaptive Generation of Privileged Intermediate Information for Visible-Infrared Person Re-Identification](https://mlsanity.com/paper/2307.03240/) (2023)
- [Improving the trustworthiness of image classification models by utilizing bounding-box annotations](https://mlsanity.com/paper/2108.10131/) (2021)
- [Neglected Free Lunch -- Learning Image Classifiers Using Annotation Byproducts](https://mlsanity.com/paper/2303.17595/) (2023)
- [Teachers in concordance for pseudo-labeling of 3D sequential data](https://mlsanity.com/paper/2207.06079/) (2022)
- [Audio Representation Learning by Distilling Video as Privileged Information](https://mlsanity.com/paper/2302.02845/) (2023)
- [Guided Policy Optimization under Partial Observability](https://mlsanity.com/paper/2505.15418/) (2025)
- [MPVNN: Mutated Pathway Visible Neural Network Architecture for Interpretable Prediction of Cancer-specific Survival Risk](https://mlsanity.com/paper/2202.00882/) (2022)
- [Rethinking Knowledge Transfer in Learning Using Privileged Information](https://mlsanity.com/paper/2408.14319/) (2024)
- [Meta-learning using privileged information for dynamics](https://mlsanity.com/paper/2104.14290/) (2021)
- [Distilling Privileged Information for Dubins Traveling Salesman Problems with Neighborhoods](https://mlsanity.com/paper/2404.16721/) (2024)
- [A Multi-View Joint Learning Framework for Embedding Clinical Codes and Text Using Graph Neural Networks](https://mlsanity.com/paper/2301.11608/) (2023)
- [Exemplar Diffusion: Improving Medical Object Detection with Opportunistic Labels](https://mlsanity.com/paper/2603.15267/) (2026)
- [Neurosymbolic Imitation Learning with Human Guidance: A Privileged Information Approach](https://mlsanity.com/paper/2605.07166/) (2026)
- [Sequential Minimal Optimization Algorithm for One-Class Support Vector Machines With Privileged Information](https://mlsanity.com/paper/2606.22210/) (2026)
- [AIDE: Automated Instruction via Distilled Expertise for Reference-Free Motor Skill Coaching](https://mlsanity.com/paper/2608.03047/) (2026)
- [Quantifying the Value of Privileged Information Using a PAC-Bayesian Approach](https://mlsanity.com/paper/2609.12891/) (2026)
- [Kinematics-Centric Continuous Sign Language Retrieval with Gloss-Guided Boundary-Aware Alignment](https://mlsanity.com/paper/2610.05306/) (2026)

## Abstract (arXiv)

Unlike machines, humans learn through rapid, abstract model-building. The role of a teacher is not simply to hammer home right or wrong answers, but rather to provide intuitive comments, comparisons, and explanations to a pupil. This is what the Learning Under Privileged Information (LUPI) paradigm endeavors to model by utilizing extra knowledge only available during training. We propose a new LUPI algorithm specifically designed for Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs). We propose to use a heteroscedastic dropout (i.e. dropout with a varying variance) and make the variance of the dropout a function of privileged information. Intuitively, this corresponds to using the privileged information to control the uncertainty of the model output. We perform experiments using CNNs and RNNs for the tasks of image classification and machine translation. Our method significantly increases the sample efficiency during learning, resulting in higher accuracy with a large margin when the number of training examples is limited. We also theoretically justify the gains in sample efficiency by providing a generalization error bound decreasing with O(frac1n), where n is the number of training examples, in an oracle case.