Device for determining network configurations of a neural network under a number of ancillary conditions

摘要

Device for the determination of a Pareto set of network configurations, identifying the network configurations for a neural network for a given application and determining the specified application in the form of training data provided, including a machine-readable storage medium, stored on the command which, when running through a computer, caused the computer to perform a procedure with the following steps:(a) provide (S1) a majority of randomly selected network configurations (a);b) training (S2) of neural networks configured with the network configurations (a) on the provided training data;c) Determine (S3) a performance (accuracy) for each of the trained neural networks;(d) evaluating (S3) each of the neural networks trained with respect to at least one specified optimization target (n-pars);(e) Adding (S4) of a tube comprehensive network configurations (a) and performance (accuracy) and the evaluated optimization goal (n-pars) to another training data set;(f) Adjust (S5) a prediction model depending on the further training data set in such a way that the prediction model is given to one of the network configurations (a) from the next training data set at least its assigned performance from the next training data set;(g) select (S6) a variety of network configurations using the prediction model so that the selected network configurations meet a predetermined criterion;(h) re-hollow steps (b) to (g) with the variety of network configurations selected until a termination criterion is met; andi) Select (S7) of the network configurations (a) from the further training data set, which corresponds to a Pareto set in terms of performance and optimization goal.