In‐Depth Analysis of One Selector–One Resistor Crossbar Array for Its Writing and Reading Operations for Hardware Neural Network with Finite Wire Resistance

Automatic environmental monitoring is a field that encompasses several scientific practices for the assessment of risks that may negatively impact a given environment, such as the forest. A forest is a natural environment that hosts various forms of plant and animal life, so preserving the forest is a top priority. To this end, the authors of this paper will focus on the development of an intelligent system for the early detection of forest fires, based on an IoT solution. This latter will thus facilitate the exploitation of the functionalities offered by the Cloud and mobile applications. Detecting and predicting forest fires with accuracy is a difficult task that requires machine learning and an in-depth analysis of environmental conditions. This leads the authors to adopt the forward neural network algorithm by highlighting its contribution through real experiments, performed on the prototype developed in this paper.

Download Full-text

User Experience Evaluation of Cross-Channel Consumption: Based on Grounded Theory and Neural Network

Wireless Communications and Mobile Computing ◽

10.1155/2021/1133414 ◽

2021 ◽

Vol 2021 ◽

pp. 1-12

Author(s):

Hua Yang ◽

Huiying Wei ◽

Xiang He ◽

Yue Yan ◽

Xiaoju Liu

Keyword(s):

Neural Network ◽

Grounded Theory ◽

User Experience ◽

Index System ◽

Evaluation Model ◽

Evaluation Index System ◽

Theoretical Research ◽

Evaluation Index ◽

Channel Switching ◽

Depth Analysis

With the rapid development of e-commerce technology, cross-channel consumption has become the mainstream mode of contemporary consumers. However, there are several problems of cross-channel consumption such as inconsistency of online and offline channel information and service, disfluency of channel switching which have brought adverse effects on user experience. The question arises here as to what factors influence user experience and how to build a scientific and effective evaluation index system. Different from previous studies based on sellers, this paper used grounded theory to analyze and summarize the evaluation index system of user experience under cross-channel consumption from the perspective of consumers. We summarized and refined four first level indexes which are “online platform attribute, offline entity attribute, channel switching attribute, and individual demand” and 13 second level indexes which are “platform operation, platform information, platform service, platform promotion, product quality, service quality, environment quality, channel consistency, channel switching cost, channel switching fluency, psychological expectation, personal interests and individual needs.” Then, we used BP neural network to build the evaluation model and trained and simulated the performance of the sample. The results show that the evaluation model has a good generalization ability and can effectively evaluate user experience under cross-channel consumption. Finally, implications and limitations are also discussed. This study helps to enrich the theoretical research on user experience and consumer behavior. It also provides targeted basis for in-depth analysis of cross-channel consumption behavior, establishment of user experience evaluation index system, and improving user experience and multichannel management of physical stores.

Download Full-text

Neural Network Training Acceleration With RRAM-Based Hybrid Synapses

Frontiers in Neuroscience ◽

10.3389/fnins.2021.690418 ◽

2021 ◽

Vol 15 ◽

Author(s):

Wooseok Choi ◽

Myonghoon Kwak ◽

Seyoung Kim ◽

Hyunsang Hwang

Keyword(s):

Neural Network ◽

Parallel Learning ◽

Network Training ◽

Multiple Devices ◽

Network Simulations ◽

Conductance States ◽

Hardware Neural Network ◽

Training Efficiency ◽

Computation Speed ◽

Intrinsic Gain

Hardware neural network (HNN) based on analog synapse array excels in accelerating parallel computations. To implement an energy-efficient HNN with high accuracy, high-precision synaptic devices and fully-parallel array operations are essential. However, existing resistive memory (RRAM) devices can represent only a finite number of conductance states. Recently, there have been attempts to compensate device nonidealities using multiple devices per weight. While there is a benefit, it is difficult to apply the existing parallel updating scheme to the synaptic units, which significantly increases updating process’s cost in terms of computation speed, energy, and complexity. Here, we propose an RRAM-based hybrid synaptic unit consisting of a “big” synapse and a “small” synapse, and a related training method. Unlike previous attempts, array-wise fully-parallel learning is possible with our proposed architecture with a simple array selection logic. To experimentally verify the hybrid synapse, we exploit Mo/TiOx RRAM, which shows promising synaptic properties and areal dependency of conductance precision. By realizing the intrinsic gain via proportionally scaled device area, we show that the big and small synapse can be implemented at the device-level without modifications to the operational scheme. Through neural network simulations, we confirm that RRAM-based hybrid synapse with the proposed learning method achieves maximum accuracy of 97 %, comparable to floating-point implementation (97.92%) of the software even with only 50 conductance states in each device. Our results promise training efficiency and inference accuracy by using existing RRAM devices.

Download Full-text

An electrical hardware neural network of an optimization-problem solver

Neural Networks ◽

10.1016/0893-6080(88)90499-6 ◽

1988 ◽

Vol 1 ◽

pp. 475

Keyword(s):

Neural Network ◽

Optimization Problem ◽

Problem Solver ◽

Hardware Neural Network

Download Full-text

Impact of Synaptic Device Variations on Pattern Recognition Accuracy in a Hardware Neural Network

Scientific Reports ◽

10.1038/s41598-018-21057-x ◽

2018 ◽

Vol 8 (1) ◽

Cited By ~ 24

Author(s):

Sungho Kim ◽

Meehyun Lim ◽

Yeamin Kim ◽

Hee-Dong Kim ◽

Sung-Jin Choi

Keyword(s):

Neural Network ◽

Pattern Recognition ◽

Recognition Accuracy ◽

Hardware Neural Network

Download Full-text

Simulation of information decoding processes in the output device of the biomorphic neuroprocessor

Tyumen State University Herald Physical and Mathematical Modeling Oil Gas Energy ◽

10.21684/2411-7978-2020-6-4-179-193 ◽

2020 ◽

Vol 6 (4) ◽

pp. 179-193

Author(s):

Alexander D. Pisarev ◽

Alexander N. Busygin ◽

Abdulla Kh. A. Ibrahim ◽

Sergey Yu. Udovichenko

Keyword(s):

Neural Network ◽

Pulse Generator ◽

Input Pulse ◽

Electrical Circuit ◽

Input Device ◽

Binary Number ◽

Information Process ◽

Output Device ◽

The Matrix ◽

Hardware Neural Network

This publication is the series of articles continuation on the creation of neuroprocessor nodes based on a composite memristor-diode crossbar. The authors have determined the principles of modifying the pulse information into a binary code in the output device of the neuroprocessor, implemented in a logical matrix based on a new electronic element — a combined memristor-diode crossbar. The processing of pulse signals is possible in the logical matrix, since one layer of the matrix is a set of logical AND or OR gates with arbitrarily connected inputs. The authors have proposed two solutions to the problem of decoding pulses from a population of neurons in the output device, coming from the hardware neural network of the neuroprocessor, into standard binary signals. The first solution involves the two layers use of a logical matrix and a pulse generator. The compactness of the second solution is achieved due to the presence of a binary number generator, which allows to get rid of one layer of the logical matrix. This article presents the SPICE modeling results of the decoding pulsed information process signals into binary format and confirms the operability of the output device electrical circuit. The originality of the device operation lies in the switching of the generator signals by the logical matrix to the neuroprocessor output based on the time delay of the input pulse from the hardware neural network. The use of the memristor logical matrix in all nodes of the neuroprocessor, including the input device, makes it possible to unify the element base of the neuroprocessor complete electrical circuit, as well as its power supplies.

Download Full-text