Highly Linear and Symmetric Weight Modification in HfO2‐Based Memristive Devices for High‐Precision Weight Entries

Advanced Electronic Materials - Tập 6 Số 9 - 2020
Jin Joo Ryu1,2, Kanghyeok Jeon2,3, Guhyun Kim3, Min Yang4, Chunjoong Kim1, Doo Seok Jeong3, Gun Hwan Kim2
1Department of Materials Science and Engineering, Chungnam National University, Daejeon, 34134, Republic of Korea
2Division of Advanced Materials, Korea Research Institute of Chemical Technology (KRICT), 141 Gajeong-ro, Yuseong-gu, Daejeon 34114, Republic of Korea
3Division of Materials Science and Engineering, Hanyang University, Wansimni-ro 222, Seongdong-gu, Seoul, 04763 Republic of Korea
4Intelligent Electronic Device Lab, Sahmyook University, 815 Hwarang-ro, Nowon-Gu, Seoul, 01795 Republic of Korea

Tóm tắt

AbstractIn this study, highly reliable and accurate weight‐modification behaviors are realized using a W/Al2O3 (3 nm)/HfO2 (7 nm)/TiN memristive device. The accuracy of the simulated inference of the MNIST dataset when considering the weight‐modification behavior is ≈95%. It is determined the optimal programming voltage pulsing conditions considering i) a high linearity in the weight‐modification, ii) symmetry between potentiation and depression, and iii) an alleviation of the voltage‐driving circuit overhead for the related part of weight‐modification process. Particular emphasis is placed on the last concern, and thus, the fixed shape of each programming pulse for both potentiation and depression are utilized. The optimal pulse design is 500 µs for the pulse rising, plateau, and falling times and a 2 V amplitude at the absolute scale. Additionally, the nonparametric method to evaluate the linearity and symmetry as opposed to the application of several parametric methods are proposed. The nonparametric method is based on an evaluation of actual data rather than models, and thus considers the actual variability in the conductance change, which is otherwise often ignored in the parameter optimization process.

Từ khóa


Tài liệu tham khảo

Krizhevsky A., 2012, Adv. Neural Inf. Process. Syst., 25, 1097

Taigman Y., 2014, IEEE Conf. Comput. Vision Pattern Recogn., 27, 1701

T.Mikolov K.Chen G.Corrado J.Dean Efficient estimation of word representation in vector space ICLR workshop arXiv:1302.37812013.

Collobert R., 2011, J. Mach. Learn. Res., 12, 2493

10.1038/nature14236

10.1038/nature16961

10.1038/nature24270

10.1038/nature14539

Hennessy J. L., 2011, Computer Architecture A Quantitative Approach

Gao M., 2015, Int. Conf. Parallel Architecture and Compilation (PACT), 113

T.vincon A.Koch I.Petrov Moving processing to data: On the influence of processing in memory on data management arXiv:1905.047672019.

Devaux F., 2019, 2019 IEEE Hot Chips 31 Symp. (HCS), 1

Okumura S., 2019, 2019 Symp. on VLSI Circuits, C248, 10.23919/VLSIC.2019.8778187

10.3389/fnins.2019.01383

10.1109/TMSCS.2017.2721160

Imani M., 2019, Proc. the 46th Int. Symp. on Computer Architecture (ISCA), 802, 10.1145/3307650.3322237

10.1002/adma.201704729

10.1038/s41578-019-0159-3

10.1109/JPROC.2012.2190369

Burr G. W., 2016, Adv. Phys.: X, 2, 89

10.1002/admt.201900037

Carver A. M., 1990, Proc. IEEE, 78, 10

10.1109/LED.2017.2719161

10.1109/TED.2016.2615648

10.1021/acsami.8b20035

10.1016/j.cplett.2004.01.080

10.1016/S0040-6090(98)00975-4

10.1016/0257-8972(95)02594-4

10.1109/JPROC.2017.2684830

10.1063/1.3355415

10.1063/1.5040430

10.1088/1361-6528/ab3480

10.1016/j.mser.2014.06.002

Thông tin
Thông tin xuất bản

Advanced Electronic Materials

Tập 6 Số 9

Thông tin tác giả