Data denoising method based on mark risk control

Abstract

The invention discloses a data denoising method based on mark risk control, and aims to solve the problem that the success of data deep learning often depends on a large amount of accurately marked data, but a large amount of accurately marked data is usually difficult to collect in an actual scene. In order to reduce the influence of data marking noise on the performance of the neural networks, the method maintains the two neural networks to mutually select data with small loss as low-risk data to update the peer-to-peer network, and each network filters out high-risk data in the network and retrains the remaining data. Along with training, the two networks are more and more similar to cause learning performance degradation, in order to solve the problem, when the inconsistency of the two neural networks is stable, mutual data selection is stopped, and the obtained low-risk data is utilized to update the networks until convergence. Compared with the prior art, the invention enables the deep neural network to have higher robustness.

Description

technical field [0001] The invention relates to a data denoising method based on marker risk control, which can screen out high-risk marker data and improve robustness, and belongs to the technical field of computer artificial intelligence data analysis. Background technique [0002] In recent years, deep learning has achieved great success in various fields, such as face recognition, automatic driving, machine translation and so on. However, the performance of deep learning depends on a large amount of accurately labeled data. In practical applications, it is often difficult to collect a large amount of accurately labeled data, because obtaining accurate data labels requires a lot of manpower and material resources. To solve this problem, people usually use crowdsourcing technology to allocate a large amount of unlabeled data to voluntary users for labeling. Due to the uneven level of users, the obtained labels are often noisy. How to learn from data with labeled noise has...

AI Technical Summary

Problems solved by technology

However, the performance of deep learning depends on a large amount of accurately labeled data. In practical applications, it is often difficult to collect a large amount of accurately labeled data, because obtaining accurate data labels requires a lot of manpower and material resources.

Images