Publications associées à Floating-point error mitigation

Functional-Basis Analysis of Non-Stationary Signals in Modern Power Grids: Theory and Implementation in Embedded Systems

Situational awareness strategies are essential for the reliable and secure operation of the electric power grid which represents critical infrastructure in modern society. With the rise of converter-interfaced renewable generation and the consequent shift ...

EPFL2024

A 16-bit Floating-Point Near-SRAM Architecture for Low-power Sparse Matrix-Vector Multiplication

David Atienza Alonso, Giovanni Ansaloni, Grégoire Axel Eggermann, Marco Antonio Rios

State-of-the-art Artificial Intelligence (AI) algorithms, such as graph neural networks and recommendation systems, require floating-point computation of very large matrix multiplications over sparse data. Their execution in resource-constrained scenarios, ...

New York2023

Efficient Lung Nodule Detection via 3D Deep Learning with Shifted Convolutions

Jiancheng Yang, Bo Du, Kang Yuan

The high computational costs of deep convolutional neural networks hinder their deployment in real-world applications, including pulmonary nodule detection from CT scans where large 3D image sizes amplify the issue. This paper presents a novel 3D method to ...

IEEE2023

Iterative refinement of Schur decompositions

Daniel Kressner, Zvonimir Bujanovic

The Schur decomposition of a square matrix A is an important intermediate step of state-of-the-art numerical algorithms for addressing eigenvalue problems, matrix functions, and matrix equations. This work is concerned with the following task: Compute a (m ...

SPRINGER2023

ColTraIn: Co-located DNN training and inference

Mario Paulo Drumond Lages De Oliveira

Deep neural network inference accelerators are deployed at scale to accommodate online services, but face low average load because of service demand variability, leading to poor resource utilization. Unfortunately, reclaiming inference idle cycles is diffi ...

EPFL2020

ESSOP: Efficient and Scalable Stochastic Outer Product Architecture for Deep Learning

Irem Boybat Kara, Evangelos Eleftheriou, Abu Sebastian

Deep neural networks (DNNs) have surpassed human-level accuracy in a variety of cognitive tasks but at the cost of significant memory/time requirements in DNN training. This limits their deployment in energy and memory limited applications that require rea ...

IEEE2020

Design of approximate and precision-scalable circuits for embedded multimedia and neural-network processing

Vincent Frédéric Camus

Density, speed and energy efficiency of integrated circuits have been increasing exponentially for the last four decades following Moore's law. However, power and reliability pose several challenges to the future of technology scaling. Approximate computin ...

EPFL2019

Multi-ReRAM synapses for artificial neural network training

Elmira Shahrabi, Igor Krawczuk, Yusuf Leblebici, Irem Boybat Kara, Carlo Ricciardi, Cecilia Giovinazzo, Evangelos Eleftheriou, Abu Sebastian

Metal-oxide-based resistive memory devices (ReRAM) are being actively researched as synaptic elements of neuromorphic co-processors for training deep neural networks (DNNs). However, device-level non-idealities are posing significant challenges. In this wo ...

IEEE2019

Advanced Machine Learning Techniques for Self-Interference Cancellation in Full-Duplex Radios

Andreas Peter Burg, Andreas Toftegaard Kristensen

In-band full-duplex systems allow for more efficient use of temporal and spectral resources by transmitting and receiving information at the same time and on the same frequency. However, this creates a strong self-interference signal at the receiver, makin ...

IEEE2019

Training DNNs with Hybrid Block Floating Point

Babak Falsafi, Martin Jaggi, Tao Lin, Mario Paulo Drumond Lages De Oliveira

The wide adoption of DNNs has given birth to unrelenting computing requirements, forcing datacenter operators to adopt domain-specific accelerators to train them. These accelerators typically employ densely packed full-precision floating-point arithmetic t ...

NEURAL INFORMATION PROCESSING SYSTEMS (NIPS)2018

A Householder-Based Algorithm For Hessenberg-Triangular Reduction

Daniel Kressner, Zvonimir Bujanovic

The QZ algorithm for computing eigenvalues and eigenvectors of a matrix pencil A - lambda B requires that the matrices first be reduced to Hessenberg-triangular (HT) form. The current method of choice for HT reduction relies entirely on Givens rotations re ...

SIAM PUBLICATIONS2018

Towards a Compiler for Reals

Viktor Kuncak, Eva Darulova

Numerical software, common in scientific computing or embedded systems, inevitably uses a finite-precision approximation of the real arithmetic in which most algorithms are designed. In many applications, the roundoff errors introduced by finite-precision ...

2017

A Low-power Carry Cut-Back Approximate Adder with Fixed-point Implementation and Floating-point Precision

Christian Enz, Vincent Frédéric Camus, Jérémy Lucien Maurice Schlachter

This paper introduces an approximate adder architecture based on a digital quasi-feedback technique called Carry Cut-Back in which high-significance stages can cut the carry propagation chain at lower-significance positions. This lightweight approach preve ...

Ieee2016

A Bio-Inspired Reconfigurable Architecture for Local Binary Descriptors

Selman Ergünay, Yusuf Leblebici

Limited resources of embedded devices and increased real time constraints have raised interest to the binary description methods over the floating-point ones such as SIFT and SURF in computer vision applications. Although many software applications of the ...

Ieee2016

JPEG XT: A Compression Standard for HDR and WCG Images [Standards in a Nutshell]

Touradj Ebrahimi, Philippe Hanhart

High bit depth data acquisition and manipulation have been largely studied at the academic level over the last 15 years and are rapidly attracting interest at the industrial level. An example of the increasing interest for high-dynamic range (HDR) imaging ...

Institute of Electrical and Electronics Engineers2016

Programming with Numerical Uncertainties

Eva Darulova

Numerical software, common in scientific computing or embedded systems, inevitably uses an approximation of the real arithmetic in which most algorithms are designed. In many domains, roundoff errors are not the only source of inaccuracy and measurement as ...

EPFL2014

Fast Directional Spatially Localized Spherical Harmonic Transform

Yves Wiaux, Jason Douglas McEwen

We propose a transform for signals defined on the sphere that reveals their localized directional content in the spatio-spectral domain when used in conjunction with an asymmetric window function. We call this transform the directional spatially localized ...

Institute of Electrical and Electronics Engineers2013

Synthesis of Fixed-Point Programs

Viktor Kuncak, Eva Darulova, Indranil Saha

Several problems in the implementations of control systems, signal-processing systems, and scientific computing systems reduce to compiling a polynomial expression over the reals into an imperative program using fixed-point arithmetic. Fixed-point arithmet ...

2013

Miniboxing: Improving the Speed to Code Size Tradeoff in Parametric Polymorphism Translations

Martin Odersky, Vlad Ureche, Cristian Talau

Parametric polymorphism enables code reuse and type safety. Underneath the uniform interface exposed to programmers, however, its low level implementation has to cope with inherently non-uniform data: value types of different sizes and semantics (bytes, in ...

2013

Thick Boundaries in Binary Space and Their Influence on Nearest-Neighbor Search

Pascal Fua, Vincent Lepetit, Tomasz Piotr Trzcinski

Binary descriptors allow faster similarity computation than real-valued ones while requiring much less storage. As a result, many algorithms have recently been proposed to binarize ﬂoating-point descriptors so that they can be searched for quickly. Unfortu ...

Elsevier2012