Publications

Results 7476–7500 of 99,299

Search results

Jump to search filters

Inverse Methods - Users Manual 5.6

Walsh, Timothy; Akcelik, Volkan; Aquino, Wilkins; Mccormick, Cameron; Sanders, Clay; Treweek, Benjamin; Kurzawski, John C.; Smith, Chandler

The inverse methods team provides a set of tools for solving inverse problems in structural dynamics and thermal physics, and also sensor placement optimization via Optimal Experimental Design (OED). These methods are used for designing experiments, model calibration, and verfication/validation analysis of weapons systems. This document provides a user's guide to the input for the three apps that are supported for these methods. Details of input specifications, output options, and optimization parameters are included.

More Details
TYPE SAND Report YEAR 2022
DOIOSTI

An Accurate, Error-Tolerant, and Energy-Efficient Neural Network Inference Engine Based on SONOS Analog Memory

IEEE Transactions on Circuits and Systems I: Regular Papers

Xiao, Tianyao P.; Feinberg, Benjamin; Bennett, Christopher; Agrawal, Vineet; Saxena, Prashant; Prabhakar, Venkatraman; Ramkumar, Krishnaswamy; Medu, Harsha; Raghavan, Vijay; Chettuvetty, Ramesh; Agarwal, Sapan; Marinella, Matthew

We demonstrate SONOS (silicon-oxide-nitride-oxide-silicon) analog memory arrays that are optimized for neural network inference. The devices are fabricated in a 40nm process and operated in the subthreshold regime for in-memory matrix multiplication. Subthreshold operation enables low conductances to be implemented with low error, which matches the typical weight distribution of neural networks, which is heavily skewed toward near-zero values. This leads to high accuracy in the presence of programming errors and process variations. We simulate the end-To-end neural network inference accuracy, accounting for the measured programming error, read noise, and retention loss in a fabricated SONOS array. Evaluated on the ImageNet dataset using ResNet50, the accuracy using a SONOS system is within 2.16% of floating-point accuracy without any retraining. The unique error properties and high On/Off ratio of the SONOS device allow scaling to large arrays without bit slicing, and enable an inference architecture that achieves 20 TOPS/W on ResNet50, a > 10× gain in energy efficiency over state-of-The-Art digital and analog inference accelerators.

More Details
TYPE Journal Article YEAR 2022
DOIOSTIScopus
Results 7476–7500 of 99,299
Results 7476–7500 of 99,299