Structured data sample increment method and device, electronic equipment and medium

Abstract

The invention discloses an increment method of a structured data sample. The objective of the invention is to solve the technical defect that an existing generative adversarial network cannot generatea new sample for an original sample of which the data type is structured data. The method comprises the following steps that: before a structured data original sample is input into a generative adversarial network, conversion operation of converting non-quantized structural data into quantized parameters is firstly executed, namely, the non-quantized structural information is expressed by using the quantized parameters, so that an original sample expressed by using the quantized parameters meets the precondition of generating a new sample according to the original sample through a generativeadversarial network. Through the technical scheme provided by the invention, the application field range of the generative adversarial network is expanded to the field of structured data, so that thestructured data sample can also generate a high-quality new sample by utilizing the generative adversarial network. The invention further discloses a structured data sample increment device, electronic equipment and a computer readable storage medium, which have the above beneficial effects.

Description

technical field [0001] The present application relates to the field of new sample generation, and in particular to a method, device, electronic device and computer-readable storage medium for incrementally constructing structured data samples. Background technique [0002] Generative Adversarial Networks (GAN, Generative Adversarial Networks) is a deep learning model and one of the most promising methods for unsupervised learning on complex distributions in recent years. Different from other deep learning models, the generative model (Generative Model) and the discriminative model (Discriminative Model) of the mutual game in the generative confrontation network can produce high-quality output. The process of the game is also a process of confrontational learning. In the game process: the generative model is responsible for generating new data based on the input data, and strives to pass the discrimination of the new data generated by the discriminant model. New data that me...

AI Technical Summary

Problems solved by technology

[0006] This application provides an incremental method, device, electronic device, and computer-readable storage medium for structured data samples, aiming to solve the problem that existing generative confrontation networks cannot be used to generate new samples whose data type is the original sample of structured data question

Images