Memristive LSTM network hardware architecture for time-series predictive modeling problems

Kazybek Adam, Kamilya Smagulova, Alex James Pappachen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Analysis of time-series data allows to identify long term trends and make predictions that can help to improve our lives. With rapid development of artificial neural networks, long short-Tern memory (LSTM) recurrent neural network (RNN) configuration is found to be capable in dealing with time-series forecasting problems where data points are time dependent and possess seasonality trends. Gated structure of LSTM cell and flexibility in network topology (one-To-many, many-To-one, etc) allows to model systems with multiple input variables and control several parameters such as the size of look-back window to make a prediction and number of time steps to be predicted. These make LSTM attractive tool over conventional methods such as auto regression models, simple average, moving average, naive approach, ARIMA, Holts linear trend method, Holts Winter seasonal method, and others. In this paper, we propose a hardware implementation of LSTM network architecture for time-series forecasting problem. Circuit simulations were performed in LTspice using TSMC 0.18 μm CMOS technology. Software simulations were performed in Python using Keras library.

Original languageEnglish
Title of host publication2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages459-462
Number of pages4
ISBN (Electronic)9781538682401
DOIs
Publication statusPublished - Jan 8 2019
Event14th IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018 - Chengdu, China
Duration: Oct 26 2018Oct 30 2018

Publication series

Name2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018

Conference

Conference14th IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018
CountryChina
CityChengdu
Period10/26/1810/30/18

Fingerprint

Computer networks
Computer hardware
Time series
hardware
Data storage equipment
Computer simulation
trends
forecasting
Recurrent neural networks
Circuit simulation
Network architecture
predictions
winter
Topology
regression analysis
Neural networks
CMOS
flexibility
topology
simulation

Keywords

  • analog circuit
  • crossbar
  • LSTM
  • memristor
  • RNN
  • time-series prediction

ASJC Scopus subject areas

  • Biomedical Engineering
  • Electrical and Electronic Engineering
  • Instrumentation

Cite this

Adam, K., Smagulova, K., & James Pappachen, A. (2019). Memristive LSTM network hardware architecture for time-series predictive modeling problems. In 2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018 (pp. 459-462). [8605649] (2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/APCCAS.2018.8605649

Memristive LSTM network hardware architecture for time-series predictive modeling problems. / Adam, Kazybek; Smagulova, Kamilya; James Pappachen, Alex.

2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018. Institute of Electrical and Electronics Engineers Inc., 2019. p. 459-462 8605649 (2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Adam, K, Smagulova, K & James Pappachen, A 2019, Memristive LSTM network hardware architecture for time-series predictive modeling problems. in 2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018., 8605649, 2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018, Institute of Electrical and Electronics Engineers Inc., pp. 459-462, 14th IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018, Chengdu, China, 10/26/18. https://doi.org/10.1109/APCCAS.2018.8605649
Adam K, Smagulova K, James Pappachen A. Memristive LSTM network hardware architecture for time-series predictive modeling problems. In 2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018. Institute of Electrical and Electronics Engineers Inc. 2019. p. 459-462. 8605649. (2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018). https://doi.org/10.1109/APCCAS.2018.8605649
Adam, Kazybek ; Smagulova, Kamilya ; James Pappachen, Alex. / Memristive LSTM network hardware architecture for time-series predictive modeling problems. 2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 459-462 (2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018).
@inproceedings{8aa24115414540c9bf49486c24ec45e2,
title = "Memristive LSTM network hardware architecture for time-series predictive modeling problems",
abstract = "Analysis of time-series data allows to identify long term trends and make predictions that can help to improve our lives. With rapid development of artificial neural networks, long short-Tern memory (LSTM) recurrent neural network (RNN) configuration is found to be capable in dealing with time-series forecasting problems where data points are time dependent and possess seasonality trends. Gated structure of LSTM cell and flexibility in network topology (one-To-many, many-To-one, etc) allows to model systems with multiple input variables and control several parameters such as the size of look-back window to make a prediction and number of time steps to be predicted. These make LSTM attractive tool over conventional methods such as auto regression models, simple average, moving average, naive approach, ARIMA, Holts linear trend method, Holts Winter seasonal method, and others. In this paper, we propose a hardware implementation of LSTM network architecture for time-series forecasting problem. Circuit simulations were performed in LTspice using TSMC 0.18 μm CMOS technology. Software simulations were performed in Python using Keras library.",
keywords = "analog circuit, crossbar, LSTM, memristor, RNN, time-series prediction",
author = "Kazybek Adam and Kamilya Smagulova and {James Pappachen}, Alex",
year = "2019",
month = "1",
day = "8",
doi = "10.1109/APCCAS.2018.8605649",
language = "English",
series = "2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "459--462",
booktitle = "2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018",
address = "United States",

}

TY - GEN

T1 - Memristive LSTM network hardware architecture for time-series predictive modeling problems

AU - Adam, Kazybek

AU - Smagulova, Kamilya

AU - James Pappachen, Alex

PY - 2019/1/8

Y1 - 2019/1/8

N2 - Analysis of time-series data allows to identify long term trends and make predictions that can help to improve our lives. With rapid development of artificial neural networks, long short-Tern memory (LSTM) recurrent neural network (RNN) configuration is found to be capable in dealing with time-series forecasting problems where data points are time dependent and possess seasonality trends. Gated structure of LSTM cell and flexibility in network topology (one-To-many, many-To-one, etc) allows to model systems with multiple input variables and control several parameters such as the size of look-back window to make a prediction and number of time steps to be predicted. These make LSTM attractive tool over conventional methods such as auto regression models, simple average, moving average, naive approach, ARIMA, Holts linear trend method, Holts Winter seasonal method, and others. In this paper, we propose a hardware implementation of LSTM network architecture for time-series forecasting problem. Circuit simulations were performed in LTspice using TSMC 0.18 μm CMOS technology. Software simulations were performed in Python using Keras library.

AB - Analysis of time-series data allows to identify long term trends and make predictions that can help to improve our lives. With rapid development of artificial neural networks, long short-Tern memory (LSTM) recurrent neural network (RNN) configuration is found to be capable in dealing with time-series forecasting problems where data points are time dependent and possess seasonality trends. Gated structure of LSTM cell and flexibility in network topology (one-To-many, many-To-one, etc) allows to model systems with multiple input variables and control several parameters such as the size of look-back window to make a prediction and number of time steps to be predicted. These make LSTM attractive tool over conventional methods such as auto regression models, simple average, moving average, naive approach, ARIMA, Holts linear trend method, Holts Winter seasonal method, and others. In this paper, we propose a hardware implementation of LSTM network architecture for time-series forecasting problem. Circuit simulations were performed in LTspice using TSMC 0.18 μm CMOS technology. Software simulations were performed in Python using Keras library.

KW - analog circuit

KW - crossbar

KW - LSTM

KW - memristor

KW - RNN

KW - time-series prediction

UR - http://www.scopus.com/inward/record.url?scp=85062219376&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062219376&partnerID=8YFLogxK

U2 - 10.1109/APCCAS.2018.8605649

DO - 10.1109/APCCAS.2018.8605649

M3 - Conference contribution

T3 - 2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018

SP - 459

EP - 462

BT - 2018 IEEE Asia Pacific Conference on Circuits and Systems, APCCAS 2018

PB - Institute of Electrical and Electronics Engineers Inc.

ER -