A METHOD AND APPARATUS FOR MACHINE LEARNING BASED ON WEAKLY SUPERVISED LEARNING

摘要

A weakly supervised learning-based machine learning method performed by a machine learning apparatus according to an embodiment of the present invention is a method in which first type information is given using convolutional neural networks, 2 type information includes extracting a feature map for an unassigned data set, performing a task corresponding to the first type information using a first model, And updating the convolution neural network by performing a task corresponding to the first type information using a second model different from the first model and back propagating the computed second error value to update the convolution neural network Wherein when performing the task corresponding to the first type information using the second model, the second type information can be extracted