A deep learning heterogeneous computing method based on layer-wide memory allocation, at least comprises steps of: traversing a neural network model so as to acquire a training operational sequence and a number of layers L thereof; calculating a memory room R.sub.1 required by data involved in operation at the i.sup.th layer of the neural network model under a double-buffer configuration, where 1≤i≤L; altering a layer structure of the i.sup.th layer and updating the training operational sequence; distributing all the data across a memory room of the CPU and the memory room of the GPU according to a data placement method; performing iterative computation at each said layer successively based on the training operational sequence so as to complete neural network training.

Disclosed is a computer-implemented method for estimating an uncertainty of a prediction generated by a machine learning system, the method including: receiving first data; training a first machine learning model component of a machine learning system with the received first data, the first machine learning model component is trained to generate a prediction; generating an uncertainty estimate of the prediction; training a second machine learning model component of the machine learning system with second data, the second machine learning model component is trained to generate a calibrated uncertainty estimate of the prediction. Also disclosed is a corresponding system.

Disclosed is a computer-implemented method for estimating an uncertainty of a prediction generated by a machine learning system, the method including: receiving first data; training a first machine learning model component of a machine learning system with the received first data, the first machine learning model component is trained to generate a prediction; generating an uncertainty estimate of the prediction; training a second machine learning model component of the machine learning system with second data, the second machine learning model component is trained to generate a calibrated uncertainty estimate of the prediction. Also disclosed is a corresponding system.

A processor-implemented method of performing convolution operations in a neural network includes generating a plurality of first sub-bit groups and a plurality of second sub-bit groups, respectively from at least one pixel value of an input feature map and at least one predetermined weight, performing a convolution operation on a first pair that includes a first sub-bit group including a most significant bit (MSB) of the at least one pixel value and a second sub-bit group including an MSB of the at least one predetermined weight, based on the plurality of second sub-bit groups, obtaining a maximum value of a sum of results for convolution operations of remaining pairs excepting the first pair, and based on a result of the convolution operation on the first pair and the maximum value, determining whether to perform the convolution operations of the remaining pairs.

A device for operating a machine learning system. The machine learning system is assigned a predefinable rollout, which characterizes a sequence in which each of the layers ascertains an intermediate variable. When assigning the rollout, each connection or each layer is assigned a control variable, which characterizes whether the intermediate variable of each of the subsequent connected layers is ascertained according to the sequence or regardless of the sequence. A calculation of an output variable of the machine learning system as a function of an input variable of the machine learning system is controlled as a function of the predefinable rollout. Also described is a method for operating the machine learning system.

A device for operating a machine learning system. The machine learning system is assigned a predefinable rollout, which characterizes a sequence in which each of the layers ascertains an intermediate variable. When assigning the rollout, each connection or each layer is assigned a control variable, which characterizes whether the intermediate variable of each of the subsequent connected layers is ascertained according to the sequence or regardless of the sequence. A calculation of an output variable of the machine learning system as a function of an input variable of the machine learning system is controlled as a function of the predefinable rollout. Also described is a method for operating the machine learning system.

Provided are a method for processing model parameters, and an apparatus. The method comprises: a model parameter set to be sharded is obtained, wherein the model parameter set comprises a multi-dimensional array corresponding to a feature embedding; attribute information for a storage system used for storing the model parameter set to be sharded is obtained, wherein the storage system used for storing the model parameter set to be sharded differs from a system on which a model corresponding to the model parameter set to be sharded is located when operating; the model parameter set to be sharded is stored in the storage system according to the attribute information.

A method for using a graph neural network framework to improve learning and predicting in a multiplex network environment is provided. The method includes: identifying a plurality of layers of a multiplex network; estimating, for each layer, a corresponding probability of selecting the layer as being a relevant layer for training with respect to an application; estimating, for each layer, a corresponding loss associated with selecting the layer as being relevant; calculating, for each layer based on the corresponding probability and the corresponding loss, a corresponding regret associated with selecting the layer as being relevant; determining, for each layer based on the calculated corresponding regret, whether to select the layer as being relevant; and training the multiplex network with respect to the application by aggregating information obtained from layers that have been determined as being relevant layers.

A method for using a graph neural network framework to improve learning and predicting in a multiplex network environment is provided. The method includes: identifying a plurality of layers of a multiplex network; estimating, for each layer, a corresponding probability of selecting the layer as being a relevant layer for training with respect to an application; estimating, for each layer, a corresponding loss associated with selecting the layer as being relevant; calculating, for each layer based on the corresponding probability and the corresponding loss, a corresponding regret associated with selecting the layer as being relevant; determining, for each layer based on the calculated corresponding regret, whether to select the layer as being relevant; and training the multiplex network with respect to the application by aggregating information obtained from layers that have been determined as being relevant layers.

A method and system is provided for identifying patterns in datasets by identifying delimited regions of feature-space in which patterns occur. The delimited regions are then combined into an ensemble able to make predictions based on the identified regions of feature-space. The method may be used for classification, for regression, for auto-encoding, for simulation, and for other applications of pattern detection.