VALSE

VALSE 首页 活动通知 查看内容

VALSE 论文速览 第87期: 双视角一致的在线增量学习

2022-7-20 17:17| 发布者: 程一-计算所| 查看: 108| 评论: 0

摘要: 为了使得视觉与学习领域相关从业者快速及时地了解领域的最新发展动态和前沿技术进展,VALSE最新推出了《论文速览》栏目,将在每周发布一至两篇顶会顶刊论文的录制视频,对单个前沿工作进行细致讲解。本期VALSE论文速 ...

为了使得视觉与学习领域相关从业者快速及时地了解领域的最新发展动态和前沿技术进展,VALSE最新推出了《论文速览》栏目,将在每周发布一至两篇顶会顶刊论文的录制视频,对单个前沿工作进行细致讲解。本期VALSE论文速览选取了来自西安电子科技大学的在线增量学习方面的工作。该工作由论文第一作者顾亚男博士录制,由邓成教授指导。


论文题目:Not Just Selection, but Exploration: 双视角一致的在线增量学习

作者列表:顾亚男 (西安电子科技大学),杨旭 (西安电子科技大学),魏坤 (西安电子科技大学),邓成 (西安电子科技大学)

B站观看网址:

https://www.bilibili.com/video/BV1rZ4y1Y7dv/


论文摘要:

Online class-incremental continual learning aims to learn new classes continually from a never-ending and single-pass data stream, while not forgetting the learned knowledge of old classes. Existing replay-based methods have shown promising performance by storing a subset of old class data. Unfortunately, these methods only focus on selecting samples from the memory bank for replay and ignore the adequate exploration of semantic information in the single-pass data stream, leading to poor classification accuracy. In this paper, we propose a novel yet effective framework for online class-incremental continual learning, which considers not only the selection of stored samples, but also the full exploration of the data stream. Specifically, we propose a gradient-based sample selection strategy, which selects the stored samples whose gradients generated in the network are most interfered by the new incoming samples. We believe such samples are beneficial for updating the neural network based on back gradient propagation. More importantly, we seek to explore the semantic information between two different views of training images by maximizing their mutual information, which is conducive to the improvement of classification accuracy. Extensive experimental results demonstrate that our method achieves state-of-the-art performance on a variety of benchmark datasets.


论文信息:

[1] Yanan Gu, Xu Yang, Kun Wei, Cheng Deng. Not Just Selection, but Exploration: Online Class-Incremental Continual Learning via Dual View Consistency. CVPR 2022.


论文链接:

[https://openaccess.thecvf.com/content/CVPR2022/papers/Gu_Not_Just_Selection_but_Exploration_Online_Class-Incremental_Continual_Learning_via_CVPR_2022_paper.pdf]


代码链接:

[https://github.com/YananGu/DVC]


视频讲者简介:

顾亚男,西安电子科技大学电子工程学院博士生,师从邓成教授,主要研究方向是计算机视觉和深度增量学习,在CVPR、AAAI等国际会议发表论文多篇。



特别鸣谢本次论文速览主要组织者:

月度轮值AC:王智慧 (大连理工大学)、杨旭 (西安电子科技大学)

季度责任AC:魏秀参 (南京理工大学)


活动参与方式

1、VALSE每周举行的Webinar活动依托B站直播平台进行,欢迎在B站搜索VALSE_Webinar关注我们!

直播地址:

https://live.bilibili.com/22300737;

历史视频观看地址:

https://space.bilibili.com/562085182/ 


2、VALSE Webinar活动通常每周三晚上20:00进行,但偶尔会因为讲者时区问题略有调整,为方便您参加活动,请关注VALSE微信公众号:valse_wechat 或加入VALSE QQ R群,群号:137634472);


*注:申请加入VALSE QQ群时需验证姓名、单位和身份缺一不可。入群后,请实名,姓名身份单位。身份:学校及科研单位人员T;企业研发I;博士D;硕士M。


3、VALSE微信公众号一般会在每周四发布下一周Webinar报告的通知。


4您也可以通过访问VALSE主页:http://valser.org/ 直接查看Webinar活动信息。Webinar报告的PPT(经讲者允许后),会在VALSE官网每期报告通知的最下方更新。

小黑屋|手机版|Archiver|Vision And Learning SEminar

GMT+8, 2022-8-13 11:07 , Processed in 0.013405 second(s), 14 queries .

Powered by Discuz! X3.4

Copyright © 2001-2020, Tencent Cloud.

返回顶部