paper-with-me

홈 › Papers

Quantized Approximately Orthogonal Recurrent Neural Networks

2024-02-05 · Armand Foucault, Franck Mamalet, François Malgouyres

In recent years, Orthogonal Recurrent Neural Networks (ORNNs) have gained popularity due to their ability to manage tasks involving long-term dependencies, such as the copy-task, and their linear complexity. However, existing ORNNs utilize full precision weights and activations, which prevents their deployment on compact devices.In this paper, we explore the quantization of the weight matrices in ORNNs, leading to Quantized approximately Orthogonal RNNs (QORNNs). The construction of such networks remained an open problem, acknowledged for its inherent instability. We propose and investigate two strategies to learn QORNN by combining quantization-aware training (QAT) and orthogonal projections. We also study post-training quantization of the activations for pure integer computation of the recurrent loop. The most efficient models achieve results similar to state-of-the-art full-precision ORNN, LSTM and FastRNN on a variety of standard benchmarks, even with 4-bits quantization.

📄 PDF Abstract BibTeX arXiv:2402.04012

Code (0)

등록된 구현이 없습니다.

Tasks

QuantizationTime Series

Methods 이 논문이 사용한 방법론

Sigmoid Activation 설명 없음
Tanh Activation 설명 없음
LSTM An LSTM is a type of recurrent neural network that addresses the vanishing gradient problem in vanilla…

Similar Papers 제목 키워드 기반

Coordinate descent on the orthogonal group for recurrent neural network training

2021-07-30 · Estelle Massart, Vinayak Abrol

We propose to use stochastic Riemannian coordinate descent on the orthogonal group for recurrent neural network training. The algorithm rotates successively two columns of the recurrent matrix, an operation that can be e…

QCM-SGM+: Improved Quantized Compressed Sensing With Score-Based Generative Models

2023-02-02 · Xiangming Meng, Yoshiyuki Kabashima

In practical compressed sensing (CS), the obtained measurements typically necessitate quantization to a limited number of bits prior to transmission or storage. This nonlinear quantization process poses significant recov…

Bayesian Inferencecompressed sensingQuantization

When Quantization Breaks Memory: Recurrent-State Write-Back in Low-Precision Temporal Inference

2026-09-03 · Ismail Erbas, Xavier Intes, Vikas Pandey hf

Quantization is widely used to reduce the computational and memory demands of neural-network inference. In recurrent networks, however, the quantized state is stored and returned at the next time step, so the rule used t…

Forget by Uncertainty: Orthogonal Entropy Unlearning for Quantized Neural Networks

2026-01-31 · Tian Zhang, Yujia Tong, Junhao Dong, Ke Xu 외 arxiv

The deployment of quantized neural networks on edge devices, combined with privacy regulations like GDPR, creates an urgent need for machine unlearning in quantized models. However, existing methods face critical challen…

Orthogonal Gated Recurrent Unit with Neumann-Cayley Transformation

2022-08-12 · Edison Mucllari, Vasily Zadorozhnyy, Cole Pospisil, Duc Nguyen 외

In recent years, using orthogonal matrices has been shown to be a promising approach in improving Recurrent Neural Networks (RNNs) with training, stability, and convergence, particularly, to control gradients. While Gate…