# Detecting Unreliable Classifications from Quantized Models

arXiv: https://arxiv.org/abs/2610.18575 · HTML: https://mlsanity.com/paper/2610.18575/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 14 references that are in the map.
- First posted on arXiv 13 Oct 2026.

## Builds on (most cited first, 10 of 10)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [A Survey on Deep Learning in Medical Image Analysis](https://mlsanity.com/paper/1702.05747/) (2017)
- [Loss Aware Post-training Quantization](https://mlsanity.com/paper/1911.07190/) (2019)
- [Training Dynamics Impact Post-Training Quantization Robustness](https://mlsanity.com/paper/2510.06213/) (2025)
- [Great Models Think Alike: Improving Model Reliability via Inter-Model Latent Agreement](https://mlsanity.com/paper/2305.01481/) (2023)
- [Cross-Model Disagreement as a Label-Free Correctness Signal](https://mlsanity.com/paper/2603.25450/) (2026)
- [Saliency Assisted Quantization for Neural Networks](https://mlsanity.com/paper/2411.05858/) (2024)
- [Uncovering Overconfident Failures in CXR Models via Augmentation-Sensitivity Risk Scoring](https://mlsanity.com/paper/2510.01683/) (2025)

## Abstract (arXiv)

Post-training INT8 quantization is typically validated by checking accuracy retention, which requires ground-truth labels often unavailable in low-resource clinical settings. We show a label-free alternative: comparing a model's own output before and after quantization. We compute cosine similarity between the FP32 and INT8 models' output probability vectors for each prediction, which requires only a single forward pass. This output-agreement similarity separates lower-accuracy predictions across four evaluated architectures (ResNet-50, VGG-16-BN, Wide ResNet-50, and ViT-B/16) under fixed-percentile splits, outperforming a confidence-only baseline for all four. These results show that quantized models can be compared to their baselines to assess their drift and determine whether they are well suited for clinical deployment.