Panasonic Intellectual Property Corporation of America

Denis A Gudovskiy

Sotaro Tsukizawa

Luca Rigazio

Takuya Yamaguchi

Yasunori Ishii

Kazuki Kozuka

Alec Hodgkinson

Shun Ishizaka

First optimization processing for optimizing parameters of a DNN and second optimization processing for optimizing hyperpararneters for each sample used in data augmentation processing are alternately performed. The first optimization processing includes causing the DNN to predict a first augmentation label from a first augmented sample, calculating a first error function between the first augmentation label and a first correct label for a first sample, and updating the parameters in accordance with the first error function. The second optimization processing includes acquiring a second sample, causing the DNN after the updating of the parameters to predict a second label from the second sample, calculating a second error function between the second label and a second correct label for the second sample, and updating the hyperparameter in accordance with a gradient obtained by differentiation of the second error function with respect to the hyperparameter.

Data augmentation method, learning device, and recording medium

An information processing method implemented by a computer includes: obtaining a piece of first data, and a piece of second data not included in a training dataset for training an inferencer; calculating, using a piece of first relevant data obtained by inputting the first data to the inferencer trained by machine learning using the training dataset, a first contribution representing contributions of portions constituting the first data to a piece of first output data output by inputting the first data to the inferencer; calculating, using a piece of second relevant data obtained by inputting the second data to the inferencer, a second contribution representing contributions of portions constituting the second data to a piece of second output data output by inputting the second data to the inferencer; and determining whether to add the second data to the training dataset, according to the similarity between the first and second contributions.

Information processing method and information processing system

Executing a deep neural network by obtaining, during deep neural network inference, a binary intermediate feature map in binary representation by converting a floating-point or fixed-point intermediate feature map into a binary vector using a first transformation module; generating a compressed feature map by compressing the binary intermediate feature map using a nonlinear dimensionality reduction layer; storing the compressed feature map into memory; reconstructing the binary intermediate feature map by decompressing the compressed feature map read from the memory using a reconstruction layer corresponding to the nonlinear dimensionality reduction layer; and converting the reconstructed binary intermediate feature map into a floating-point or fixed-point intermediate feature map using a second transformation module.

Execution method, execution device, learning method, learning device, and recording medium for deep neural network

An information processing method performed by a computer includes: obtaining a plurality of recognition result candidates in sensing data and a likelihood of each of the plurality of recognition result candidates, the plurality of recognition result candidates and the likelihood being obtained by inputting the sensing data to a model that is trained by machine learning and performs recognition processing; obtaining an indication designating a part to be analyzed in the sensing data; selecting at least one recognition result candidate from the plurality of recognition result candidates, based on (i) a relationship between each of the plurality of recognition result candidates and the part and (ii) the likelihood of each of the plurality of recognition result, candidates; and outputting the at least one recognition result candidate that is selected.

An information processing method includes: inputting an input tensor indicating data to a processor having a memory; causing the processor to perform, after elements of the input tensor are subjected to precomputation for conversion into a power-of-two format and are stored in the memory, convolution operation processing with only addition and shift operations by using the precomputed elements of the input tensor stored in the memory and weight tensors that are pre-converted into the power-of-two format in accordance with a predetermined algorithm, that are stored in the memory, and that indicate weights having a possibility of being used for a convolution operation; and outputting, as an output tensor, the elements of the input tensor on which the convolution operation processing is performed.

Growing community of inventors

Emeryville, CA, United States of America

Denis A Gudovskiy