# You Never Know: Quantization Induces Inconsistent Biases in Vision-Language Foundation Models

arXiv: https://arxiv.org/abs/2410.20265 · HTML: https://mlsanity.com/paper/2410.20265/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Social bias and fairness in VLMs, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 28 of its 35 references that are in the map.
- First posted on arXiv 26 Oct 2024.

## Placed in

- Multimodal LLMs / vision-language models / [Social bias and fairness in VLMs](https://mlsanity.com/mllm/social-bias-and-fairness-in-vlms/)

## Builds on (most cited first, 28 of 28)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [LAION-5B: An open large-scale dataset for training next generation image-text models](https://mlsanity.com/paper/2210.08402/) (2022)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [A Survey of Quantization Methods for Efficient Neural Network Inference](https://mlsanity.com/paper/2103.13630/) (2021)
- [DataComp: In search of the next generation of multimodal datasets](https://mlsanity.com/paper/2304.14108/) (2023)
- [8-bit Optimizers via Block-wise Quantization](https://mlsanity.com/paper/2110.02861/) (2021)
- [What Do Compressed Deep Neural Networks Forget?](https://mlsanity.com/paper/1911.05248/) (2019)
- [Characterising Bias in Compressed Models](https://mlsanity.com/paper/2010.03058/) (2020)
- [Fairness-Aware Ranking in Search & Recommendation Systems with Application to LinkedIn Talent Search](https://mlsanity.com/paper/1905.01989/) (2019)
- [A Prompt Array Keeps the Bias Away: Debiasing Vision-Language Models with Adversarial Learning](https://mlsanity.com/paper/2203.11933/) (2022)
- [DeAR: Debiasing Vision-Language Models with Additive Residuals](https://mlsanity.com/paper/2303.10431/) (2023)
- [FairPrune: Achieving Fairness Through Pruning for Dermatological Disease Diagnosis](https://mlsanity.com/paper/2203.02110/) (2022)
- [FACET: Fairness in Computer Vision Evaluation Benchmark](https://mlsanity.com/paper/2309.00035/) (2023)
- [Pruning has a disparate impact on model accuracy](https://mlsanity.com/paper/2205.13574/) (2022)
- [Arbitrariness and Social Prediction: The Confounding Role of Variance in Fair Classification](https://mlsanity.com/paper/2301.11562/) (2023)
- [FairGRAPE: Fairness-aware GRAdient Pruning mEthod for Face Attribute Classification](https://mlsanity.com/paper/2207.10888/) (2022)
- [Gender Bias Amplification During Speed-Quality Optimization in Neural Machine Translation](https://mlsanity.com/paper/2106.00169/) (2021)
- [Qu-ANTI-zation: Exploiting Quantization Artifacts for Achieving Adversarial Outcomes](https://mlsanity.com/paper/2110.13541/) (2021)
- [ECoFLaP: Efficient Coarse-to-Fine Layer-Wise Pruning for Vision-Language Models](https://mlsanity.com/paper/2310.02998/) (2023)
- [Stealthy Backdoors as Compression Artifacts](https://mlsanity.com/paper/2104.15129/) (2021)
- [Intriguing Properties of Compression on Multilingual Models](https://mlsanity.com/paper/2211.02738/) (2022)
- [The Effect of Model Compression on Fairness in Facial Expression Recognition](https://mlsanity.com/paper/2201.01709/) (2022)
- [FairDeDup: Detecting and Mitigating Vision-Language Fairness Disparities in Semantic Dataset Deduplication](https://mlsanity.com/paper/2404.16123/) (2024)
- [Debiased Distillation by Transplanting the Last Layer](https://mlsanity.com/paper/2302.11187/) (2023)

## Built on by (most cited first, 2 of 2)

- [Less Precise Can Be More Reliable: A Systematic Evaluation of Quantization's Impact on VLMs Beyond Accuracy](https://mlsanity.com/paper/2509.21173/) (2025)
- [Edge Reliability Gap in Vision-Language Models: Quantifying Failure Modes of Compressed VLMs Under Visual Corruption](https://mlsanity.com/paper/2603.26769/) (2026)

## Abstract (arXiv)

We study the impact of a standard practice in compressing foundation vision-language models - quantization - on the models' ability to produce socially-fair outputs. In contrast to prior findings with unimodal models that compression consistently amplifies social biases, our extensive evaluation of four quantization settings across three datasets and three CLIP variants yields a surprising result: while individual models demonstrate bias, we find no consistent change in bias magnitude or direction across a population of compressed models due to quantization.