# Teaching AI to Teach: Leveraging Limited Human Salience Data Into Unlimited Saliency-Based Training

arXiv: https://arxiv.org/abs/2306.05527 · HTML: https://mlsanity.com/paper/2306.05527/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Placed in General knowledge distillation methods, one of 12 directions in Knowledge distillation.
- Builds on 25 of its 35 references that are in the map.
- First posted on arXiv 8 Jun 2023.

## Placed in

- Knowledge distillation / [General knowledge distillation methods](https://mlsanity.com/distill/general-knowledge-distillation-methods/)

## Builds on (most cited first, 25 of 25)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [Densely Connected Convolutional Networks](https://mlsanity.com/paper/1608.06993/) (2016)
- [Rethinking the Inception Architecture for Computer Vision](https://mlsanity.com/paper/1512.00567/) (2015)
- [Analyzing and Improving the Image Quality of StyleGAN](https://mlsanity.com/paper/1912.04958/) (2019)
- [Progressive Growing of GANs for Improved Quality, Stability, and Variation](https://mlsanity.com/paper/1710.10196/) (2017)
- [Learning Deep Features for Discriminative Localization](https://mlsanity.com/paper/1512.04150/) (2015)
- [Xception: Deep Learning with Depthwise Separable Convolutions](https://mlsanity.com/paper/1610.02357/) (2016)
- [Training Generative Adversarial Networks with Limited Data](https://mlsanity.com/paper/2006.06676/) (2020)
- [Alias-Free Generative Adversarial Networks](https://mlsanity.com/paper/2106.12423/) (2021)
- [StarGAN v2: Diverse Image Synthesis for Multiple Domains](https://mlsanity.com/paper/1912.01865/) (2019)
- [Grad-CAM++: Improved Visual Explanations for Deep Convolutional Networks](https://mlsanity.com/paper/1710.11063/) (2017)
- [RISE: Randomized Input Sampling for Explanation of Black-box Models](https://mlsanity.com/paper/1806.07421/) (2018)
- [Score-CAM: Score-Weighted Visual Explanations for Convolutional Neural Networks](https://mlsanity.com/paper/1910.01279/) (2019)
- [Eigen-CAM: Class Activation Map using Principal Components](https://mlsanity.com/paper/2008.00299/) (2020)
- [img2pose: Face Alignment and Detection via 6DoF, Face Pose Estimation](https://mlsanity.com/paper/2012.07791/) (2020)
- [CYBORG: Blending Human Saliency Into the Loss Improves Deep Learning](https://mlsanity.com/paper/2112.00686/) (2021)
- [Iris Liveness Detection Competition (LivDet-Iris) -- The 2020 Edition](https://mlsanity.com/paper/2009.00749/) (2020)
- [Human-Aided Saliency Maps Improve Generalization of Deep Learning](https://mlsanity.com/paper/2105.03492/) (2021)
- [Domain-Specific Human-Inspired Binarized Statistical Image Features for Iris Recognition](https://mlsanity.com/paper/1807.05248/) (2018)
- [SREFI: Synthesis of Realistic Example Face Images](https://mlsanity.com/paper/1704.06693/) (2017)
- [Unfooling Perturbation-Based Post Hoc Explainers](https://mlsanity.com/paper/2205.14772/) (2022)
- [Human Saliency-Driven Patch-based Matching for Interpretable Post-mortem Iris Recognition](https://mlsanity.com/paper/2208.03138/) (2022)
- [Measuring Human Perception to Improve Open Set Recognition](https://mlsanity.com/paper/2209.03519/) (2022)
- [Using Human Perception to Regularize Transfer Learning](https://mlsanity.com/paper/2211.07885/) (2022)

## Built on by (most cited first, 8 of 8)

- [MENTOR: Human Perception-Guided Pretraining for Increased Generalization](https://mlsanity.com/paper/2310.19545/) (2023)
- [Taking Training Seriously: Human Guidance and Management-Based Regulation of Artificial Intelligence](https://mlsanity.com/paper/2402.08466/) (2024)
- [Increasing Interpretability of Neural Networks By Approximating Human Visual Saliency](https://mlsanity.com/paper/2410.16115/) (2024)
- [Saliency-Guided Training for Fingerprint Presentation Attack Detection](https://mlsanity.com/paper/2505.02176/) (2025)
- [Psych-Occlusion: Using Visual Psychophysics for Aerial Detection of Occluded Persons during Search and Rescue](https://mlsanity.com/paper/2412.05553/) (2024)
- [Almost Right: Making First-Layer Kernels Nearly Orthogonal Improves Model Generalization](https://mlsanity.com/paper/2504.16362/) (2025)
- [VISER: Visually-Informed System for Enhanced Robustness in Iris Presentation Attack Detection](https://mlsanity.com/paper/2603.17859/) (2026)
- [What's Old is New Again: Classical Dimensionality Reduction for Efficient Saliency-Guided Biometric Attack Detection](https://mlsanity.com/paper/2606.13528/) (2026)

## Abstract (arXiv)

Machine learning models have shown increased accuracy in classification tasks when the training process incorporates human perceptual information. However, a challenge in training human-guided models is the cost associated with collecting image annotations for human salience. Collecting annotation data for all images in a large training set can be prohibitively expensive. In this work, we utilize "teacher" models (trained on a small amount of human-annotated data) to annotate additional data by means of teacher models' saliency maps. Then, "student" models are trained using the larger amount of annotated training data. This approach makes it possible to supplement a limited number of human-supplied annotations with an arbitrarily large number of model-generated image annotations. We compare the accuracy achieved by our teacher-student training paradigm with (1) training using all available human salience annotations, and (2) using all available training data without human salience annotations. We use synthetic face detection and fake iris detection as example challenging problems, and report results across four model architectures (DenseNet, ResNet, Xception, and Inception), and two saliency estimation methods (CAM and RISE). Results show that our teacher-student training paradigm results in models that significantly exceed the performance of both baselines, demonstrating that our approach can usefully leverage a small amount of human annotations to generate salience maps for an arbitrary amount of additional training data.