US 11,816,553 B2
Output from a recurrent neural network
Henry Markram, Pully (CH); Felix Schürmann, Grens (CH); Fabien Jonathan Delalondre, Geneva (CH); Daniel Milan Lütgehetmann, Lausanne (CH); John Rahmon, Lausanne (CH); Constantin Cosmin Atanasoaei, Chavannes-pres-Renens (CH); and Michele De Gruttola, Geneva (CH)
Assigned to INAIT SA, Lausanne (CH)
Filed by INAIT SA, Lausanne (CH)
Filed on Dec. 11, 2019, as Appl. No. 16/710,205.
Prior Publication US 2021/0182653 A1, Jun. 17, 2021
Int. Cl. G06N 3/02 (2006.01); G06N 3/044 (2023.01); G06N 3/04 (2023.01); G06N 3/08 (2023.01); G06N 3/082 (2023.01)
CPC G06N 3/044 (2023.01) [G06N 3/02 (2013.01); G06N 3/04 (2013.01); G06N 3/08 (2013.01); G06N 3/082 (2013.01)] 27 Claims
OG exemplary drawing
 
17. A neural network system comprising:
an artificial recurrent neural network implemented by data processing apparatus, wherein the artificial recurrent neural network is configured to preprocess input data and comprises
a first region that is configured to receive first data, wherein the first data comprises a first class data included in the input data,
a second region that is configured to receive second data, wherein the second data comprises a second class of data included in the input data that is different from the first class of data, the first region is primarily perturbed by the first data and the second region is primarily perturbed by the second data even when both regions are perturbed at the same time, and
a third region that is configured to receive results of processing by both the first region and by the second regions, wherein the third region is configured to output a representation of the input data that is suitable for input into multiple diverse applications, the representation comprising indications of the presence of topological patterns of signal transmission activity that occurs amongst groups of three or more nodes in the artificial recurrent neural network, wherein at least one of the topological patterns of signal transmission activity represents a respective abstraction of a same characteristic that is present in multiple types of data included in the input data and the topological patterns are indicative of the results of the processing by the first region and by the second region; and
a first application coupled to receive at least some of the indications of the presence of topological patterns of signal transmission activity.