US 11,704,561 B2
Method for realizing a neural network
Thomas Hinterstoisser, Bisamberg (AT); Martin Matschnig, Tulin (AT); and Herbert Taucher, Vienna (AT)
Assigned to SIEMENS AKTIENGESELLSCHAFT, Munich (DE)
Appl. No. 16/955,356
Filed by Siemens Aktiengesellschaft, Munich (DE)
PCT Filed Dec. 12, 2018, PCT No. PCT/EP2018/084487
§ 371(c)(1), (2) Date Jun. 18, 2020,
PCT Pub. No. WO2019/121206, PCT Pub. Date Jun. 27, 2019.
Claims priority of application No. 17208770 (EP), filed on Dec. 20, 2017.
Prior Publication US 2021/0097388 A1, Apr. 1, 2021
Int. Cl. G06F 30/323 (2020.01); G06F 30/327 (2020.01); G06N 3/08 (2023.01); G06N 3/063 (2023.01)
CPC G06N 3/08 (2013.01) [G06F 30/323 (2020.01); G06F 30/327 (2020.01); G06N 3/063 (2013.01)] 7 Claims
OG exemplary drawing
 
1. A method for realizing an artificial neural network via an electronic integrated circuit, the artificial neural network being formed from artificial neurons which are grouped into different layers and linked to each other, the method comprising:
a. creating a functional description, taking into account a specifiable starting weighting for each neuron;
b. performing a synthesis for each respective neuron based on a respective functional description with an associated specifiable starting weighting;
c. creating a netlist as a synthesis result, at least one base element and a starting configuration belonging to the base element being stored in the netlist for each neuron, and the at least one base element being formed by a lookup table (LUT) unit and an associated dynamic configuration cell, in which a respective current configuration for the associated LUT unit is stored;
d. implementing the netlist as a starting configuration of the artificial neural network in the electronic integrated circuit;
wherein, starting from the starting configuration of the artificial neural network implemented in the electronic integrated circuit, a training phase of the artificial neural network is performed in which at least one of (i) the starting configuration and (ii) a respective current configuration of at least one of (i) at least one base element and (ii) at least one neuron is changed; and
wherein at least one of (i) fixed, specified test data and (ii) test samples are utilized in the training phase of the artificial neural network, output data obtained with at least one of (i) the test data and (ii) test sample is compared with specified reference data, and a change to the respective current configuration of at least one of (i) at least one base unit and (ii) at least one neuron is performed iteratively until the output data obtained with at least one of (i) the test data and (ii) the specified test sample corresponds to the specified reference data within a specifiable tolerance.