Learning static bound management parameters for analog resistive processing unit system

A computer program product, comprising:

• one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions comprising:
• program instructions to perform a first training process to train a first artificial neural network model;
• program instructions to perform a second training process to retrain the first artificial neural network model using matrix-vector compute operations which are a function of bound management parameters of an analog resistive processing unit system, to thereby generate a second artificial neural network model with learned static bound management parameters;
• and
• program instructions to configure the analog resistive processing unit system to implement the second artificial neural network model and the learned static bound management parameters.

A neuromorphic computing system, comprising:

• a resistive processing unit system configured to implement a trained artificial neural network;
• wherein the resistive processing unit system comprises bound management circuitry configured to apply a static input vector scaling parameter to digital input vectors that are input to a synaptic weight matrix of at least one layer of the trained artificial neural network, and to apply a static output vector scaling parameter to digital output vectors that are generated by performing matrix-vector operations on the synaptic weight matrix of the at least one layer of the trained artificial neural network;
• and
• wherein the static input vector scaling parameter and the static output vector scaling parameter are learned parameters that are learned as part of a training process performed to train the artificial neural network implemented by the resistive processing unit system.