Self-paced resistance learning against overfitting on noisy labels

نویسندگان

چکیده

• We propose a novel resistance loss to alleviate model overfitting framework robustly train CNNs on noisy labels It integrates CNNs’ memorization effect, curriculum learning with resis-tance Noisy composed of correct and corrupted ones are pervasive in practice. They might significantly deteriorate the performance convolutional neural networks (CNNs), because easily overfitted labels. To address this issue, inspired by an observation, deep first memorize probably correct-label data then corrupt-label samples, we yet simple self-paced resist labels, without using any clean validation data. The proposed utilizes effect learn curriculum, which contains confident samples provides meaningful supervision for other training samples. Then it adopts selected update parameters; tends smooth parameters’ or attain equivalent prediction over each class, thereby resisting Finally, unify these two modules into single function optimize alternative learning. Extensive experiments demonstrate superior recent state-of-the-art methods noisy-label Source codes method available https://github.com/xsshi2015/Self-paced-Resistance-Learning .

برای دانلود باید عضویت طلایی داشته باشید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Learning with Noisy Labels

In this paper, we theoretically study the problem of binary classification in the presence of random classification noise — the learner, instead of seeing the true labels, sees labels that have independently been flipped with some small probability. Moreover, random label noise is class-conditional — the flip probability depends on the class. We provide two approaches to suitably modify any giv...

متن کامل

Self-Paced Curriculum Learning

Curriculum learning (CL) or self-paced learning (SPL) represents a recently proposed learning regime inspired by the learning process of humans and animals that gradually proceeds from easy to more complex samples in training. The two methods share a similar conceptual learning paradigm, but differ in specific learning schemes. In CL, the curriculum is predetermined by prior knowledge, and rema...

متن کامل

On the effectiveness of self-paced learning.

Metacognitive monitoring and control must be accurate and efficient in order to allow self-guided learners to improve their performance. Yet few examples exist in which allowing learners to control learning produces higher levels of performance than restricting learners' control. Here we investigate the consequences of allowing learners to self-pace study of a list of words on later recognition...

متن کامل

Learning to Tag using Noisy Labels

In order to organize and retrieve the ever growing collection of multimedia objects on the Web, many algorithms have been developed to automatically tag images, music and videos. One source of labeled data for training these algorithms are tags collected from the Web, via collaborative tagging websites (e.g., Flickr, Last.FM and YouTube) or crowdsourcing applications (e.g., human computation ga...

متن کامل

Self-Paced Learning with Diversity

Self-paced learning (SPL) is a recently proposed learning regime inspired by the learning process of humans and animals that gradually incorporates easy to more complex samples into training. Existing methods are limited in that they ignore an important aspect in learning: diversity. To incorporate this information, we propose an approach called self-paced learning with diversity (SPLD) which f...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

ژورنال

عنوان ژورنال: Pattern Recognition

سال: 2023

ISSN: ['1873-5142', '0031-3203']

DOI: https://doi.org/10.1016/j.patcog.2022.109080