Publications associées à Decimal floating point

Towards General-Purpose Decentralized Computing with Permissionless Extensibility

Smart contracts have emerged as the most promising foundations for applications of the blockchain technology. Even though smart contracts are expected to serve as the backbone of the next-generation web, they have several limitations that hinder their wide ...

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

Dynamically crossing diabolic points while encircling exceptional curves: A programmable symmetric-asymmetric multimode switch

Fabrizio Minganti

Nontrivial spectral properties of non-Hermitian systems can lead to intriguing effects with no counterparts in Hermitian systems. For instance, in a two-mode photonic system, by dynamically winding around an exceptional point (EP) a controlled asymmetric-s ...

Nature Portfolio2023

A Newton Frank-Wolfe method for constrained self-concordant minimization

Volkan Cevher, Quoc Tran Dinh

We develop a new Newton Frank-Wolfe algorithm to solve a class of constrained self-concordant minimization problems using linear minimization oracles (LMO). Unlike L-smooth convex functions, where the Lipschitz continuity of the objective gradient holds gl ...

SPRINGER2022

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

Mixed-Precision Deep Learning Based on Computational Memory

Irem Boybat Kara, Evangelos Eleftheriou, Abu Sebastian

Deep neural networks (DNNs) have revolutionized the field of artificial intelligence and have achieved unprecedented success in cognitive tasks such as image and speech recognition. Training of large DNNs, however, is computationally intensive and this has ...

FRONTIERS MEDIA SA2020

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

The adaptive interpolation method: a simple scheme to prove replica formulas in Bayesian inference

Nicolas Macris, Jean François Emmanuel Barbier

In recent years important progress has been achieved towards proving the validity of the replica predictions for the (asymptotic) mutual information (or free energy) in Bayesian inference problems. The proof techniques that have emerged appear to be quite ...

SPRINGER HEIDELBERG2019

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

Non-perturbative probability distribution function for cosmological counts in cells

Sergey Sibiryakov, Mikhail Ivanov

We present a non-perturbative calculation of the 1-point probability distribution function (PDF) for the spherically-averaged matter density field. The PDF is represented as a path integral and is evaluated using the saddle-point method. It factorizes into ...

IOP PUBLISHING LTD2019

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

Impact of Memory Voltage Scaling on Accuracy and Resilience of Deep Learning Based Edge Devices

David Atienza Alonso, Andreas Peter Burg, Miguel Peon Quiros, Andrea Bonetti, Martino Ruggiero, Benoît Walter Denkinger, Flavio Ponzina, Soumya Subhra Basu, Szabolcs Balási

Energy consumption is a significant obstacle to integrate deep learning into edge devices. Two common techniques to curve it are quantization, which reduces the size of the memories (static energy) and the number of accesses (dynamic energy), and voltage s ...

2020

Genius loci e memorie urbane. Il progetto di Ragnar Östberg per l’area di Eldkvarn

Chiara Monterumisi

The paper explores the city of Stockholm, apparently far from the main tracks of other significant European examples due to its peculiar morphological features. The study investigates the Stockholm Town Hall (Stadshuset) and the nearby never-to-be realized ...

2018

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

Logistic regression over encrypted data from fully homomorphic encryption

Zhicong Huang

Background: One of the tasks in the 2017 iDASH secure genome analysis competition was to enable training of logistic regression models over encrypted genomic data. More precisely, given a list of approximately 1500 patient records, each with 18 binary feat ...

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

Mixed-precision architecture based on computational memory for training deep neural networks

Irem Boybat Kara, Evangelos Eleftheriou, Abu Sebastian

Deep neural networks (DNN) have revolutionized the field of machine learning by providing unprecedented human-like performance in solving many real-world problems such as image or speech recognition. Training of large DNNs, however, is a computationally in ...

IEEE2018

Giant multiphoton absorption for THz resonances in silicon hydrogenic donors

Gabriel Aeppli, Le Nguyen Hoang

The absorption of multiple photons when there is no resonant intermediate state is a well-known nonlinear process in atomic vapours, dyes and semiconductors. The N-photon absorption (NPA) rate for donors in semiconductors scales proportionally from hydroge ...

Nature Publishing Group2018

Exogenous priority rules for the capacitated passenger assignment problem

Michel Bierlaire, Mohammad Yousef Maknoon, Stefan Yves Gabriel Binder

We propose a novel algorithm for the capacitated passenger assignment problem in public transportation where exogenous priority lists define the order in which passengers are assigned. Separating explicitly theses rules from the assignment procedure allows ...

Pergamon-Elsevier Science Ltd2017