International Business Machines Corporation

Robert Shelby

Ho-Cheol Kim

Geoffrey W Burr

Philip M Rice

Irem Boybat Kara

Angela-Kristina Speidel

Leslie E Krupp

Carmelo Di Nolfo

Artificial neural networks (ANNs) are a distributed computing model in which computation is accomplished using many simple processing units (called neurons) and the data embodied by the connections between neurons (called synapses) and the strength of these connections (called synaptic weights). An attractive implementation of ANNs uses the conductance of non-volatile memory (NVM) elements to code the synaptic weight. In this application, the non-idealities in the response of the NVM (such as nonlinearity, saturation, stochasticity and asymmetry in response to programming pulses) lead to reduced network performance compared to an ideal network implementation. Disclosed is a method that improves performance by implementing a learning rate parameter that is local to each synaptic connection, a method for tuning this local learning rate, and an implementation that does not compromise the ability to train many synaptic weights in parallel during learning.

Neuron-centric local learning rate for artificial neural networks to increase performance, learning rate margin, and reduce power consumption

A method of preparing a fiber including electro-spinning onto a substrate polymer solutions from a plurality of jets to form a network of filaments, wherein at least one jet sprays onto the substrate a first chemical mixture including a carbon fiber precursor compound, and at least one other jet sprays onto the substrate a second chemical mixture comprising a sacrificial polymer and a precursor compound of a functional material; and processing the filaments on the substrate, thereby forming an arrangement of carbon fibers having the functional material deposited thereon.

Methods for preparing carbon hybrid materials

Growing community of inventors

Boulder Creek, CA, United States of America

Robert Shelby