%0 Journal Article %1 moon2019temporal %A Moon, John %A Ma, Wen %A Shin, Jong Hoon %A Cai, Fuxi %A Du, Chao %A Lee, Seung Hwan %A Lu, Wei D. %D 2019 %J Nature Electronics %K learning memristor neural reservoir temporal %N 10 %P 480--487 %R 10.1038/s41928-019-0313-3 %T Temporal data classification and forecasting using a memristor-based reservoir computing system %U https://doi.org/10.1038/s41928-019-0313-3 %V 2 %X Time-series analysis including forecasting is essential in a range of fields from finance to engineering. However, long-term forecasting is difficult, particularly for cases where the underlying models and parameters are complex and unknown. Neural networks can effectively process features in temporal units and are attractive for such purposes. Reservoir computing, in particular, can offer efficient temporal processing of recurrent neural networks with a low training cost, and is thus well suited to time-series analysis and forecasting tasks. Here, we report a reservoir computing hardware system based on dynamic tungsten oxide (WOx) memristors that can efficiently process temporal data. The internal short-term memory effects of the WOx memristors allow the memristor-based reservoir to nonlinearly map temporal inputs into reservoir states, where the projected features can be readily processed by a linear readout function. We use the system to experimentally demonstrate two standard benchmarking tasks: isolated spoken-digit recognition with partial inputs, and chaotic system forecasting. A high classification accuracy of 99.2% is obtained for spoken-digit recognition, and autonomous chaotic time-series forecasting has been demonstrated over the long term.

@article{moon2019temporal, abstract = {Time-series analysis including forecasting is essential in a range of fields from finance to engineering. However, long-term forecasting is difficult, particularly for cases where the underlying models and parameters are complex and unknown. Neural networks can effectively process features in temporal units and are attractive for such purposes. Reservoir computing, in particular, can offer efficient temporal processing of recurrent neural networks with a low training cost, and is thus well suited to time-series analysis and forecasting tasks. Here, we report a reservoir computing hardware system based on dynamic tungsten oxide (WOx) memristors that can efficiently process temporal data. The internal short-term memory effects of the WOx memristors allow the memristor-based reservoir to nonlinearly map temporal inputs into reservoir states, where the projected features can be readily processed by a linear readout function. We use the system to experimentally demonstrate two standard benchmarking tasks: isolated spoken-digit recognition with partial inputs, and chaotic system forecasting. A high classification accuracy of 99.2% is obtained for spoken-digit recognition, and autonomous chaotic time-series forecasting has been demonstrated over the long term.}, added-at = {2020-05-03T14:01:56.000+0200}, author = {Moon, John and Ma, Wen and Shin, Jong Hoon and Cai, Fuxi and Du, Chao and Lee, Seung Hwan and Lu, Wei D.}, biburl = {https://www.bibsonomy.org/bibtex/25d0ac6a13f79fc1acf83dbf86008b56d/sigoldberg1}, description = {Temporal data classification and forecasting using a memristor-based reservoir computing system | Nature Electronics}, doi = {10.1038/s41928-019-0313-3}, interhash = {7e9d112312cfe5e4b40cd83d47cea118}, intrahash = {5d0ac6a13f79fc1acf83dbf86008b56d}, issn = {25201131}, journal = {Nature Electronics}, keywords = {learning memristor neural reservoir temporal}, number = 10, pages = {480--487}, refid = {Moon2019}, timestamp = {2020-05-03T14:01:56.000+0200}, title = {Temporal data classification and forecasting using a memristor-based reservoir computing system}, url = {https://doi.org/10.1038/s41928-019-0313-3}, volume = 2, year = 2019 }