# Reconstructing Training Data from Multiclass Neural Networks

arXiv: https://arxiv.org/abs/2305.03350 · HTML: https://mlsanity.com/paper/2305.03350/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 13 of its 21 references that are in the map.
- First posted on arXiv 5 May 2023.

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

- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification](https://mlsanity.com/paper/1502.01852/) (2015)
- [Extracting Training Data from Large Language Models](https://mlsanity.com/paper/2012.07805/) (2020)
- [The Secret Sharer: Evaluating and Testing Unintended Memorization in Neural Networks](https://mlsanity.com/paper/1802.08232/) (2018)
- [Extracting Training Data from Diffusion Models](https://mlsanity.com/paper/2301.13188/) (2023)
- [The Implicit Bias of Gradient Descent on Separable Data](https://mlsanity.com/paper/1710.10345/) (2017)
- [Exploiting Unintended Feature Leakage in Collaborative Learning](https://mlsanity.com/paper/1805.04049/) (2018)
- [Deep Models Under the GAN: Information Leakage from Collaborative Deep Learning](https://mlsanity.com/paper/1702.07464/) (2017)
- [Gradient Descent Maximizes the Margin of Homogeneous Neural Networks](https://mlsanity.com/paper/1906.05890/) (2019)
- [Evaluating Gradient Inversion Attacks and Defenses in Federated Learning](https://mlsanity.com/paper/2112.00059/) (2021)
- [Directional convergence and alignment in deep learning](https://mlsanity.com/paper/2006.06657/) (2020)
- [Reconstructing Training Data with Informed Adversaries](https://mlsanity.com/paper/2201.04845/) (2022)
- [Reconstructing Training Data from Trained Neural Networks](https://mlsanity.com/paper/2206.07758/) (2022)

## Built on by (most cited first, 7 of 7)

- [Provable Privacy Attacks on Trained Shallow Neural Networks](https://mlsanity.com/paper/2410.07632/) (2024)
- [Shortcut Learning Susceptibility in Vision Classifiers](https://mlsanity.com/paper/2502.09150/) (2025)
- [Fantastic DNN Classifiers and How to Identify them without Data](https://mlsanity.com/paper/2305.15563/) (2023)
- [Network Inversion for Training-Like Data Reconstruction](https://mlsanity.com/paper/2410.16884/) (2024)
- [Securing AI Systems: A Guide to Known Attacks and Impacts](https://mlsanity.com/paper/2506.23296/) (2025)
- [Novel Deep Neural Network Classifier Characterization Metrics with Applications to Dataless Evaluation](https://mlsanity.com/paper/2407.13000/) (2024)
- [Network Inversion and Its Applications](https://mlsanity.com/paper/2411.17777/) (2024)

## Abstract (arXiv)

Reconstructing samples from the training set of trained neural networks is a major privacy concern. Haim et al. (2022) recently showed that it is possible to reconstruct training samples from neural network binary classifiers, based on theoretical results about the implicit bias of gradient methods. In this work, we present several improvements and new insights over this previous work. As our main improvement, we show that training-data reconstruction is possible in the multi-class setting and that the reconstruction quality is even higher than in the case of binary classification. Moreover, we show that using weight-decay during training increases the vulnerability to sample reconstruction. Finally, while in the previous work the training set was of size at most 1000 from 10 classes, we show preliminary evidence of the ability to reconstruct from a model trained on 5000 samples from 100 classes.