US 11,704,539 B2
Forecasting routines utilizing a mixer to combine deep neural network (DNN) forecasts of multi-variate time-series datasets
Maryam Amiri, Ottawa (CA); Petar Djukic, Ottawa (CA); and Todd Morris, Stittsville (CA)
Assigned to Ciena Corporation, Hanover, MD (US)
Filed by Ciena Corporation, Hanover, MD (US)
Filed on Mar. 30, 2020, as Appl. No. 16/833,781.
Prior Publication US 2021/0303969 A1, Sep. 30, 2021
Int. Cl. G10L 15/32 (2013.01); G06N 3/045 (2023.01); G06N 3/08 (2023.01)
CPC G06N 3/045 (2023.01) [G06N 3/08 (2013.01); G10L 15/32 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A non-transitory computer-readable medium configured to store computer logic having instructions that, when executed, cause one or more processing devices to
receive, at each of a plurality of Deep Neural Network (DNN) forecasters, an input corresponding at least partially to a time-series dataset of a plurality of input time-series datasets obtained from an optical network,
produce, from each of the plurality of DNN forecasters, a forecast output,
provide the forecast output from each of the plurality of DNN forecasters to a respective input of a plurality of mixer inputs of a DNN mixer for combining the forecast outputs to produce one or more output time-series datasets,
wherein the plurality of DNN forecasters and the DNN mixer are configured in software and are executable by the one or more processing devices, and
utilize the one or more output time-series datasets to perform one or more of
predicting performance of data packets on multiple links,
predicting when a link will exceed a capacity threshold,
predicting Signal to Noise Ratio (SNR) on a fiber of an optical network,
predicting long-term quality factor of a wave, and
predicting Signal to Interference-plus-Noise Ratio (SINR) of waves on an optical submarine link.