CPC H04W 24/10 (2013.01) [G06N 3/04 (2013.01)] | 19 Claims |
1. A method of a user equipment (UE), the method comprising:
receiving an indication indicating an artificial intelligence based framework;
receiving configuration information corresponding to the artificial intelligence based framework, wherein the configuration information comprises at least two of:
a scheme corresponding to the artificial intelligence based framework, wherein the scheme comprises a convolutional neural network, a recurrent neural network, a modular neural network, or a combination thereof;
a regularization technique corresponding to the artificial intelligence based framework, wherein the regularization technique comprises a Lasso regression technique, a Ridge regression technique, a dropout technique, or a combination thereof;
a set of neural network based parameters corresponding to the artificial intelligence based framework comprising: a number of layers, a number of nodes per hidden layer, a number of input nodes, a number of output nodes, a maximum number of edges having a weight that can be reported, or a combination thereof; and
an activation function corresponding to the artificial intelligence based framework, wherein the activation function comprises a first sigmoid function, a second sigmoid function, a rectified linear unit, an arc tangent function, or a combination thereof; and
communicating an artificial intelligence report corresponding to the artificial intelligence based framework based on the configuration information, wherein the artificial intelligence report comprises a set of values corresponding to the configuration information, or an indication of a subset of a set of channel resources, or both, and wherein the artificial intelligence report comprises values corresponding to a set of neural network based parameters corresponding to the artificial intelligence based framework including: a number of inner layers, a number of hidden layers, a number of input nodes, a number of output nodes, a number of nodes per inner layer, a weight per edge between two nodes in consecutive layers, a bias per node, or a combination thereof.
|