BayesQ: Uncertainty-Guided Bayesian Quantization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lamaakal, Ismail, Yahyati, Chaymae, Maleh, Yassine, Makkaoui, Khalid El, Ouahbi, Ibrahim
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911261039525888
author Lamaakal, Ismail
Yahyati, Chaymae
Maleh, Yassine
Makkaoui, Khalid El
Ouahbi, Ibrahim
author_facet Lamaakal, Ismail
Yahyati, Chaymae
Maleh, Yassine
Makkaoui, Khalid El
Ouahbi, Ibrahim
contents We present BayesQ, an uncertainty-guided post-training quantization framework that is the first to optimize quantization under the posterior expected loss. BayesQ fits a lightweight Gaussian posterior over weights (diagonal Laplace by default; optional K-FAC/low-rank), whitens by the posterior covariance, designs codebooks to minimize posterior-expected distortion, and allocates mixed precision via a greedy knapsack that maximizes marginal expected-loss reduction per bit under a global budget. For scalar quantizers, posterior-expected MSE yields closed-form tables; task-aware proxies are handled by short Monte Carlo on a small calibration set. An optional calibration-only distillation aligns the quantized model with the posterior predictive teacher. At matched average bits/weight of 3.0/3.5/4.0, BayesQ improves over strong PTQ baselines on ResNet-50 (ImageNet) and BERT-base (GLUE) e.g., vs. GPTQ by $+1.5/+0.7/+0.3$ top-1 percentage points on RN50 and $+1.1/+0.4/+0.2$ GLUE points on BERT, while requiring one-time preprocessing comparable to a GPTQ pass. BayesQ reframes low-bit quantization as uncertainty-aware risk minimization in a practical, post-training pipeline.
format Preprint
id arxiv_https___arxiv_org_abs_2511_08821
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle BayesQ: Uncertainty-Guided Bayesian Quantization
Lamaakal, Ismail
Yahyati, Chaymae
Maleh, Yassine
Makkaoui, Khalid El
Ouahbi, Ibrahim
Machine Learning
Computation and Language
Computer Vision and Pattern Recognition
We present BayesQ, an uncertainty-guided post-training quantization framework that is the first to optimize quantization under the posterior expected loss. BayesQ fits a lightweight Gaussian posterior over weights (diagonal Laplace by default; optional K-FAC/low-rank), whitens by the posterior covariance, designs codebooks to minimize posterior-expected distortion, and allocates mixed precision via a greedy knapsack that maximizes marginal expected-loss reduction per bit under a global budget. For scalar quantizers, posterior-expected MSE yields closed-form tables; task-aware proxies are handled by short Monte Carlo on a small calibration set. An optional calibration-only distillation aligns the quantized model with the posterior predictive teacher. At matched average bits/weight of 3.0/3.5/4.0, BayesQ improves over strong PTQ baselines on ResNet-50 (ImageNet) and BERT-base (GLUE) e.g., vs. GPTQ by $+1.5/+0.7/+0.3$ top-1 percentage points on RN50 and $+1.1/+0.4/+0.2$ GLUE points on BERT, while requiring one-time preprocessing comparable to a GPTQ pass. BayesQ reframes low-bit quantization as uncertainty-aware risk minimization in a practical, post-training pipeline.
title BayesQ: Uncertainty-Guided Bayesian Quantization
topic Machine Learning
Computation and Language
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.08821