Zobrazeno 1 - 10
of 178
pro vyhledávání: '"Saab, Rayan"'
Several recent studies have investigated low-precision accumulation, reporting improvements in throughput, power, and area across various platforms. However, the accompanying proposals have only considered the quantization-aware training (QAT) paradi
Externí odkaz:
http://arxiv.org/abs/2409.17092
Autor:
Zhang, Jinjie, Saab, Rayan
Quantization is a widely used compression method that effectively reduces redundancies in over-parameterized neural networks. However, existing quantization techniques for deep neural networks often lack a comprehensive error analysis due to the pres
Externí odkaz:
http://arxiv.org/abs/2309.10975
Autor:
Maly, Johannes, Saab, Rayan
In this short note, we propose a new method for quantizing the weights of a fully trained neural network. A simple deterministic pre-processing step allows us to quantize network layers via memoryless scalar quantization while preserving the network
Externí odkaz:
http://arxiv.org/abs/2209.03487
The method of Common Spatial Patterns (CSP) is widely used for feature extraction of electroencephalography (EEG) data, such as in motor imagery brain-computer interface (BCI) systems. It is a data-driven method estimating a set of spatial filters so
Externí odkaz:
http://arxiv.org/abs/2202.04542
While neural networks have been remarkably successful in a wide array of applications, implementing them in resource-constrained hardware remains an area of intense research. By replacing the weights of a neural network with quantized (e.g., 4-bit, o
Externí odkaz:
http://arxiv.org/abs/2201.11113
We study a geometric property related to spherical hyperplane tessellations in $\mathbb{R}^{d}$. We first consider a fixed $x$ on the Euclidean sphere and tessellations with $M \gg d$ hyperplanes passing through the origin having normal vectors distr
Externí odkaz:
http://arxiv.org/abs/2108.13523
We propose the use of low bit-depth Sigma-Delta and distributed noise-shaping methods for quantizing the Random Fourier features (RFFs) associated with shift-invariant kernels. We prove that our quantized RFFs -- even in the case of $1$-bit quantizat
Externí odkaz:
http://arxiv.org/abs/2106.02614
Let $\| A \|_{\max} := \max_{i,j} |A_{i,j}|$ denote the maximum magnitude of entries of a given matrix $A$. In this paper we show that $$\max \left\{ \|U_r \|_{\max},\|V_r\|_{\max} \right\} \le \frac{(Cr)^{6r}}{\sqrt{N}},$$ where $U_r$ and $V_r$ are
Externí odkaz:
http://arxiv.org/abs/2103.13419
Autor:
Lybrand, Eric, Saab, Rayan
We propose a new computationally efficient method for quantizing the weights of pre- trained neural networks that is general enough to handle both multi-layer perceptrons and convolutional neural networks. Our method deterministically quantizes layer
Externí odkaz:
http://arxiv.org/abs/2010.15979
Autor:
Zhang, Jinjie, Saab, Rayan
We propose a fast, distance-preserving, binary embedding algorithm to transform a high-dimensional dataset $\mathcal{T}\subseteq\mathbb{R}^n$ into binary sequences in the cube $\{\pm 1\}^m$. When $\mathcal{T}$ consists of well-spread (i.e., non-spars
Externí odkaz:
http://arxiv.org/abs/2010.00712