A probabilistic framework for dynamic quantization

2025-05-15 · Gabriele Santini, Francesco Paissan, Elisabetta Farella

We propose a probabilistic framework for dynamic quantization of neural networks that allows for a computationally efficient input-adaptive rescaling of the quantization parameters. Our framework applies a probabilistic model to the network's pre-activations through a lightweight surrogate, enabling the adaptive adjustment of the quantization parameters on a per-input basis without significant memory overhead. We validate our approach on a set of popular computer vision tasks and models, observing only a negligible loss in performance. Our method strikes the best performance and computational overhead tradeoff compared to standard quantization strategies.

📄 PDF Abstract BibTeX arXiv:2505.10689

Code (0)

등록된 구현이 없습니다.

Tasks

Quantization

Methods 이 논문이 사용한 방법론

SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Deep Probabilistic Unfolding for Quantized Compressive Sensing

2026-05-12 · Gang Qu, Ping Wang, Siming Zheng, Xin Yuan arxiv

We propose a deep probabilistic unfolding model to address the classical quantized compressive sensing problem that leverages an unfolding framework to enhance the reconstruction accuracy and efficiency. Unlike previous …

Compressive Sensing

Probabilistic Learning Vector Quantization on Manifold of Symmetric Positive Definite Matrices

2021-02-01 · Fengzhen Tang, Haifeng Feng, Peter Tino, Bailu Si 외

In this paper, we develop a new classification method for manifold-valued data in the framework of probabilistic learning vector quantization. In many classification scenarios, the data can be naturally represented by sy…

EEG Electroencephalogram (EEG)General Classification Motor Imagery+1

DPQuant: Efficient and Differentially-Private Model Training via Dynamic Quantization Scheduling

2025-09-03 · Yubo Gao, Renbo Tu, Gennady Pekhimenko, Nandita Vijaykumar arxiv

Differentially-Private SGD (DP-SGD) and its adaptive variant DP-Adam are powerful techniques to protect user privacy when using sensitive data to train neural networks. During training, converting model weights and activ…

DLLMQuant: Quantizing Diffusion-based Large Language Models

2025-08-14 · Chen Xu, Dawei Yang arxiv

Diffusion-based large language models (DLLMs) have shown promise for non-autoregressive text generation, but their deployment is constrained by large model sizes and heavy computational costs. Post-training quantization …

Text Generation

Quantamination: Dynamic Quantization Leaks Your Data Across the Batch

2026-04-29 · Hanna Foerster, Ilia Shumailov, Cheng Zhang, Yiren Zhao 외 arxiv

Dynamic quantization emerged as a practical approach to increase the utilization and efficiency of the machine learning serving flow. Unlike static quantization, which applies quantization offline, dynamic quantization o…